[ccp4bb]
Chita, I disagree that the age limitation for an entry level job does anything good to a society. It gives a clear advantage to graduates from a few prestigious universities, because less fortunate students would need more time to develop their careers, no matter how talented they are. So, a competition for those positions shifts to a younger age of application to colleges, where social inequity defines a success even stronger. Moreover, it creates conditions for a corruption, because people will try to place their children in those colleges at any cost. Very poor countries cannot copy free market models from the West. Their governments should develop special programs giving an opportunity to talented kids from socially disadvantageous families to compete for good jobs. I hope your government does not require a proof of birth in a city where an institute is located? There were countries with such preconditions. Best regards, Alex On Oct 30, 2012, at 10:25 AM, Chittaranjan Das wrote: I agree with Garib that we should stop this, because nothing productive seems to be coming out of this discussion. I however like to clarify one thing about the comment I made about agism. I merely intended to interpret what the age limit preference means in the context of Government of India's policy (therefore, I used the quotation mark). I DO NOT endorse the policy. The government stipulates an age limit for any government job, including the job at MBU, IISC (last time I knew, it used to be controlled by the Government, at least in the recruitment and retirement part). Clearly, it does not confirm to the Equal Employment Opportunity (EEO) in the US, very much like many regulations in China or in Europe do not mean anything in the US. What works for US does not necessarily work for India. I always wondered what the age limit means. Why should there be one? Maybe the answer lies in the socio-economic structure of the country. Here is a country with nearly a billion people with much less opportunity (the GDP is nowhere comparable to that of US). The question for the law makers is how to distribute a very few job among the most qualified people. If we are looking at two candidates that have very similar credentials, maybe the one who accomplished them at an earlier age is better?. I think that's what the word 'preferably' signifies. I also think that it is somehow related to some grant agencies in the US determining who is a 'young investigator'. Again, because the opportunities are so few. Am I wrong? Once again, I love the Equal Employment Opportunity (EEO) in US and that's one of the reasons I came to US. But, do I think a discussion over CCP4bb can change government of India' policy and India's socio-economic status? I think not, but then, I may be a pessimist. Best Chitta - Original Message - From: Garib N Murshudov ga...@mrc-lmb.cam.ac.uk To: CCP4BB@JISCMAIL.AC.UK Sent: Tuesday, October 30, 2012 5:40:23 AM Subject: [ccp4bb] Dear all Could we stop at this point. regards Garib On 30 Oct 2012, at 18:35, Kavyashree Manjunath wrote: Dear Sir, I agree to that. But I presume that this is a platform to discuss scientific problems and not a forum to discuss or pour out personal frustrations. There may be other channels for such grievances but not this. I was just hoping this does not become a social network wherein everyone are free to express anything. Thank you kavya -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Dear Kavya, given the test is written in proper English with proper grammer etc., I think you are actually asking for censorship. I am happy ccp4bb does not censor such emails, be they in accordance with one's opinion or not! Best, Tim On 10/30/2012 06:15 AM, Kavyashree Manjunath wrote: Dear CCP4 users, It is extremely sad that CCP4BB has failed to moderate/screen for such spam mails! Thanks kavya Dear Friends, There is no need to apply to this position, we suggest. It is a PREDETERMINED SELECTION, i.e. candidate is fixed and this (advertisement, screening, selection board, selection and approval) is just the procedure. It does not matter whether you apply or not. If you apply and called for interview, then you have to waste your valuable time as well as huge travel money unless some Big Boss is fixing you to the post. Interestingly Indian Institute of Science recruits and carries faculties and trains them in such a way that it has become a epicentre of
[ccp4bb] Poor baculovirus stability at 4C?
Sorry for an off-topic question. We began experiencing a sudden reduction in stability of baculovirus stock stored at 4C, like the titer drops more than 5 fold in 3-5 month. The only difference compared with previous preparations is a switch from Gibco SF900-II to a media from Lonza (Insect-XPRESS). Could it be due to the media switch, or something else? What is the best way to store baculovirus? Thank you Alexander Aleshin Sanford-Burnham Medical Research Institute Infectious Inflammatory Disease Center 10901 North Torrey Pines Road La Jolla, California 92037
Re: [ccp4bb] offtopic: AKTA prime
it is not cheap - probably $500 When we purchased the pump, it was below $300 (a year ago). Replacement took ~4 hours, but you must like doing it. You'll have to disassemble almost entire pump module, so make pictures of each step and mark tubings ends with labels. Alex On Jul 12, 2012, at 6:28 PM, Jacqueline Vitali wrote: Yes, it has happened to me more than once. It is not a good pump. Acta Prime Plus has a better one. They still have parts. Call GE technical support and they will tell you what to do. The part costs some money (it is not cheap - probably $500). I think the Akta Prime will keep on running as long as you maintain it. They can come to your lab to do the job for you but I do not know how much it is per hour. This team is called labcrew and they are very good. I would call them and try to arrange with them. Make sure you have your solutions elevated above the pump. Even though it is a pump you need to help it. Also do not use viscous solutions with it. (I do not use glycerol because of the pump). Finally clean it up with 20% ethanol after each use to avoid mold. On Thu, Jul 12, 2012 at 5:09 PM, Peter Hsu hsuu...@u.washington.edu wrote: Hi all, We've got an old Akta prime that I think is on the verge of kicking it. Hearing some high pitched sounds coming from the pump when we're running it. Line A seems clogged and makes a thudding noise when we try to do a pump wash through that line. Does anyone have any experience w/fixing one/replacing the pump? Or any idea how much it's going to set us back to get it fixed? Sorry for the off topic question/thanks in advance for any thoughts and ideas. Peter
Re: [ccp4bb] offtopic: AKTA prime
I did not replace the entire pump, only a motor. Sorry for a confusion. On Jul 12, 2012, at 9:28 PM, aaleshin wrote: it is not cheap - probably $500 When we purchased the pump, it was below $300 (a year ago). Replacement took ~4 hours, but you must like doing it. You'll have to disassemble almost entire pump module, so make pictures of each step and mark tubings ends with labels. Alex On Jul 12, 2012, at 6:28 PM, Jacqueline Vitali wrote: Yes, it has happened to me more than once. It is not a good pump. Acta Prime Plus has a better one. They still have parts. Call GE technical support and they will tell you what to do. The part costs some money (it is not cheap - probably $500). I think the Akta Prime will keep on running as long as you maintain it. They can come to your lab to do the job for you but I do not know how much it is per hour. This team is called labcrew and they are very good. I would call them and try to arrange with them. Make sure you have your solutions elevated above the pump. Even though it is a pump you need to help it. Also do not use viscous solutions with it. (I do not use glycerol because of the pump). Finally clean it up with 20% ethanol after each use to avoid mold. On Thu, Jul 12, 2012 at 5:09 PM, Peter Hsu hsuu...@u.washington.edu wrote: Hi all, We've got an old Akta prime that I think is on the verge of kicking it. Hearing some high pitched sounds coming from the pump when we're running it. Line A seems clogged and makes a thudding noise when we try to do a pump wash through that line. Does anyone have any experience w/fixing one/replacing the pump? Or any idea how much it's going to set us back to get it fixed? Sorry for the off topic question/thanks in advance for any thoughts and ideas. Peter
Re: [ccp4bb] Fun Question - Is multiple isomorphous replacement an obsolete technique?
I wonder if anyone attempted to write a historic book on development of crystallography. That generation of crystallographers is leaving this world and soon nobody will be able to say how the protein and non-protein structures were solved in those days. Alex On Jun 6, 2012, at 8:48 AM, Gerard Bricogne wrote: Dear Fred, May I join Phil Evans in trying to dissipate the feeling that anomalous differences were fictional before flash-freezing and all the mod cons. I can remember cutting my teeth as a PhD student by helping Alan Wonacott with the experimental phasing of his B.St. GAPDH structure in 1973-74. The data were collected at room temperature on a rotating-anode source, using film on an Arndt-Wonacott rotation camera (the original prototype!). The films were scanned on a precursor of the Optronics scanner, and the intensities were integrated and scaled with the early versions of the Rotavata and Agrovata programs (mention of which should make many ccp4 old-timers swoon with nostalgia). Even with such primitive techniques, I can remember an HgI4 derivative in which you could safely refine the anomalous occupancies (i.e. f values) for the iodine atoms of the beautiful planar HgI3 anion to 5 electrons. This contributed very substantially to the phasing of the structure. In fact it would be a healthy exercise to RTFL (Read The Fascinating Literature) in this area, in particular the beautiful 1966 papers by Brian Matthews in Acta Cryst. vol 20, to see how seriously anomalous scattering was already taken as a source of phase information in macromolecular crystallography in the 1960's. In spite of that, of course, there would always be the unhappy cases where the anomalous differences were too noisy, or the data processing program too unsophisticated to filter them adequately, so that only the isomorphous differences would be useful. It was in order to carry out such filtering that Brian Matthews made another crucial contribution in the form of the Local Scaling method (Acta Cryst. A31, 480-487). With best wishes, Gerard. -- On Wed, Jun 06, 2012 at 11:02:05AM -0400, Dyda wrote: I suspect that pure MIR (without anomalous) was always a fiction. I doubt that anyone has ever used it. Heavy atoms always give an anomalous signal Phil I suspect that there was a time when the anomalous signal in data sets was fictional. Before the invent of flash freezing, systematic errors due to decay and the need of scaling together many derivative data sets collected on multiple crystals could render weak anomalous signal useless. Therefore MIR was needed. Also, current hardware/software produces much better reduced data, so weak signals can become useful. Fred [32m*** Fred Dyda, Ph.D. Phone:301-402-4496 Laboratory of Molecular BiologyFax: 301-496-0201 DHHS/NIH/NIDDK e-mail:fred.d...@nih.gov Bldg. 5. Room 303 Bethesda, MD 20892-0560 URGENT message e-mail: 2022476...@mms.att.net Google maps coords: 39.000597, -77.102102 http://www2.niddk.nih.gov/NIDDKLabs/IntramuralFaculty/DydaFred ***[m -- === * * * Gerard Bricogne g...@globalphasing.com * * * * Global Phasing Ltd. * * Sheraton House, Castle Park Tel: +44-(0)1223-353033 * * Cambridge CB3 0AX, UK Fax: +44-(0)1223-366889 * * * ===
Re: [ccp4bb] Fun Question - Is multiple isomorphous replacement an obsolete technique?
I and Victor Lamzin solved our first protein structure (3A resolution) in 80-s using pure MIR and a home made (Russian) diffractometer... Alex On Jun 6, 2012, at 1:42 PM, Boaz Shaanan wrote: So if get the gist of the thread right, am I correct in assuming that the last protein structures to be solved strictly by MIR are haemoglobin/myoglobin, lysozyme and chymotrypsin and perhaps one or two more in the late sixties? In which case the answer to the original question about MIR being obsolete, is yes it is since a long time? Boaz Boaz Shaanan, Ph.D. Dept. of Life Sciences Ben-Gurion University of the Negev Beer-Sheva 84105 Israel E-mail: bshaa...@bgu.ac.il Phone: 972-8-647-2220 Skype: boaz.shaanan Fax: 972-8-647-2992 or 972-8-646-1710 From: CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] on behalf of Phil Evans [p...@mrc-lmb.cam.ac.uk] Sent: Wednesday, June 06, 2012 6:04 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Fun Question - Is multiple isomorphous replacement an obsolete technique? No they were not useless! I used them (probably better now with cryo data though) Phil On 6 Jun 2012, at 16:02, Dyda wrote: I suspect that pure MIR (without anomalous) was always a fiction. I doubt that anyone has ever used it. Heavy atoms always give an anomalous signal Phil I suspect that there was a time when the anomalous signal in data sets was fictional. Before the invent of flash freezing, systematic errors due to decay and the need of scaling together many derivative data sets collected on multiple crystals could render weak anomalous signal useless. Therefore MIR was needed. Also, current hardware/software produces much better reduced data, so weak signals can become useful. Fred ?[32m*** Fred Dyda, Ph.D. Phone:301-402-4496 Laboratory of Molecular BiologyFax: 301-496-0201 DHHS/NIH/NIDDK e-mail:fred.d...@nih.gov Bldg. 5. Room 303 Bethesda, MD 20892-0560 URGENT message e-mail: 2022476...@mms.att.net Google maps coords: 39.000597, -77.102102 http://www2.niddk.nih.gov/NIDDKLabs/IntramuralFaculty/DydaFred ***?[m
Re: [ccp4bb] @Ian:Death of Rmerge
Wow, it is quite a lecture here! It is very appreciated. I admit some (most?) of my statements were questionable. Thus, I did not know how sigI would be calculated in case of multiple observations, and, indeed, its proper handling should make sigI/I similar to Rmerge. Consequently, I/sigI substitutes Rmerge fairly well. Now, where the metric Rmerge=0.5 came from? If I remember correctly, It was proposed here at ccp4bb. Also, one reviewer suggested to use it. I admit that this is quite an arbitrary value, but when everyone follows it, structures become comparable by this metric. If there is a better approach to estimate the resolution, lets use it, but the common rule should be enforced, otherwise the resolution becomes another venue for cheating. Once again, I was talking about metric for the resolution, it does not need to be equal to metric for the data cutoff. Alex On Jun 3, 2012, at 2:55 PM, Ian Tickle wrote: Hi Alex On 3 June 2012 07:00, aaleshin aales...@burnham.org wrote: I was also taught that under normal conditions this would occur when the data are collected up to the shell, in which Rmerge = 0.5. Do you have a reference for that? I have not seen a demonstration of such an exact relationship between Rmerge and resolution, even for 'normal' data, and I don't think everyone uses 0.5 as the cut-off anyway (e.g. some people use 0.4, some 0.8 etc - though I agree with Phil that we shouldn't get too hung up about the exact number!). Certainly having used the other suggested criteria for resolution cut-off (I/sigma(I) CC(1/2)), the corresponding Rmerge (and Rpim etc) seems to vary a lot (or maybe my data weren't 'normal'). One can collect more data (up to Rmerge=1.0 or even 100) but the resolution of the electron density map will not change significantly. I think we are all at least agreed that beyond some resolution cut-off, adding further higher resolution 'data' will not result in any further improvement in the map (because the weights will become negligible). So it would appear prudent at least to err on the high resolution side! I solved several structures of my own, and this simple rule worked every time. In what sense do you mean it 'worked'? Do you mean you tried different cut-offs in Rmerge (e.g. 0.25, 0.50, 0.75, 1.00 ...) and then used some metric to judge when there was no further significant change in the map and you noted that the optimal value of your chosen metric always occurs around Rmerge 0.5?; and if so how did you judge a 'significant change'? Personally I go along with Dale's suggestion to use the optical resolution of the map to judge when no further improvement occurs. This would need to be done with the completely refined structure because presumably optical resolution will be reduced by phase errors. Note that it wouldn't be necessary to actually quote the optical resolution in place of the X-ray resolution (that would confuse everyone!), you just need to know the value of the X-ray resolution cut-off where the optical resolution no longer changes (it should be clear from a plot of X-ray vs. optical resolution). I is measured as a number of detector counts in the reflection minus background counts. sigI is measured as sq. root of I plus standard deviation (SD) for the background plus various deviations from ideal experiment (like noise from satellite crystals). The most important contribution to the sigma(I)'s, except maybe for the weak reflections, actually comes from differences between the intensities of equivalent reflections, due to variations in absorption and illuminated volume, and other errors in image scale factors (though these are all highly correlated). These are of course exactly the same differences that contribute to Rmerge. E.g. in Scala the SDFAC SDADD parameters are automatically adjusted to fit the observed QQ plot to the expected one, in order to account for such differences. Obviously, sigI cannot be measured accurately. Moreover, the 'resolution' is related to errors in the structural factors, which are average from several measurements. Errors in their scaling would affect the 'resolution', and I/sigI does not detect them, but Rmerge does! Sorry you've lost me here, I don't see why I/sigI should not detect scaling errors: as indicated above if there are errors in the scale factors this will inflate the sigma(I) values via increased SDFAC and/or SDADD, which will increase the sigma(I) values which will in turn reduce the I/sigma(I) values exactly as expected. I see no difference in the behaviour of Rmerge and I/sigma(I) (or indeed in CC(1/2)) in this respect, since they all depend on the differences between equivalents. Rmerge, it means that the symmetry related reflections did not merge well. Under those conditions, Rmerge becomes a much better criterion for estimation of the 'resolution' than sigi/I. As indicated
Re: [ccp4bb] @Tommi:Death of Rmerge
i thought this was clear long ago.. so i am bit amazed that this discussio is still alive and kicking.. Even though it does not relate to me, but the explanation may come from the fact that new people start using crystallography, and they do not like reading old papers. So, there is nothing wrong in bringing such fundamental discussions up to life periodically. I had some misconception about accuracy of sigI, which I explained earlier. It is obvious that I/σ is the signal to noise of your measurements at a local point of the reciprocal space, but it is not obvious that it would work as well for the merged data. Thanks to Ian and others, I now understand that there is no problem with it. I am also glad to stumble on the referenced papers about the resolution and data quality (Ed Pozharski's post). I missed them because at the time when they published, I was learning molecular biology, enzymology, virology... A modern crystallographer needs to be a good biologist, and this applies some limitations on how much we know about each technique that we use. Alex On Jun 4, 2012, at 1:48 AM, Tommi Kajander wrote: well, actually i recommend having a look at the old but good scalepack manual for why Rmerge is inferior.. (i thought this was clear long ago.. so i am bit amazed that this discussio is still alive and kicking..) question of where to cut, is a different one and thats where the recent papers and developments start to come in. short quote...(scalepack manual): From a statistical point of view, I/σ is a superior criterion, for two reasons. First, it defines a resolution “limit” since by definition I/σ is the signal to noise of your measurements. In contrast, Rmerge is not directly related to signal to noise. Second, the σ assigned to each intensity derives its validity from the χ2’s, which represent the weighted ratio of the difference between the observed and average value of I, 〈I〉, squared, divided by the square of the error model, the whole thing times a factor correcting for the correlation between I and 〈I〉. Since it depends on an explicit declaration of the expected error in the measurement, the user of the program is part of the Bayesian reasoning process behind the error estimation. ..In short, I/σ is the preferred way of assessing the quality of diffraction data because it derives its validity from the χ2 (likelihood) analysis. credits to Otwinowski et al. end of story, i believe. so R-merge died long back. -tommi On Jun 4, 2012, at 9:00 AM, aaleshin wrote: Wow, it is quite a lecture here! It is very appreciated. I admit some (most?) of my statements were questionable. Thus, I did not know how sigI would be calculated in case of multiple observations, and, indeed, its proper handling should make sigI/I similar to Rmerge. Consequently, I/sigI substitutes Rmerge fairly well. Now, where the metric Rmerge=0.5 came from? If I remember correctly, It was proposed here at ccp4bb. Also, one reviewer suggested to use it. I admit that this is quite an arbitrary value, but when everyone follows it, structures become comparable by this metric. If there is a better approach to estimate the resolution, lets use it, but the common rule should be enforced, otherwise the resolution becomes another venue for cheating. Once again, I was talking about metric for the resolution, it does not need to be equal to metric for the data cutoff. Alex On Jun 3, 2012, at 2:55 PM, Ian Tickle wrote: Hi Alex On 3 June 2012 07:00, aaleshin aales...@burnham.org wrote: I was also taught that under normal conditions this would occur when the data are collected up to the shell, in which Rmerge = 0.5. Do you have a reference for that? I have not seen a demonstration of such an exact relationship between Rmerge and resolution, even for 'normal' data, and I don't think everyone uses 0.5 as the cut-off anyway (e.g. some people use 0.4, some 0.8 etc - though I agree with Phil that we shouldn't get too hung up about the exact number!). Certainly having used the other suggested criteria for resolution cut-off (I/sigma(I) CC(1/2)), the corresponding Rmerge (and Rpim etc) seems to vary a lot (or maybe my data weren't 'normal'). One can collect more data (up to Rmerge=1.0 or even 100) but the resolution of the electron density map will not change significantly. I think we are all at least agreed that beyond some resolution cut-off, adding further higher resolution 'data' will not result in any further improvement in the map (because the weights will become negligible). So it would appear prudent at least to err on the high resolution side! I solved several structures of my own, and this simple rule worked every time. In what sense do you mean it 'worked'? Do you mean you tried different cut-offs in Rmerge (e.g. 0.25, 0.50, 0.75, 1.00 ...) and then used some metric
Re: [ccp4bb] @Phil:Death of Rmerge
Could you please give me a reference to the K D paper? Without reading it, I do not see a problem with Rmerge going to infinity in high resolution shells. Indeed, I was taught at school that the crystallographic resolution is defined as a minimal distance between two peaks that can be distinguished in the electron density map. I was also taught that under normal conditions this would occur when the data are collected up to the shell, in which Rmerge = 0.5. One can collect more data (up to Rmerge=1.0 or even 100) but the resolution of the electron density map will not change significantly. I solved several structures of my own, and this simple rule worked every time. It failed only when the diffraction was very anisotropic, because the resolution was not uniform in all directions. But this obstacle can be easily overcome by presenting the resolution as a tensor with eigenvalues defined in the same simple rule (Rmerge = 0.5). Now, why such a simple method for estimation of the data resolution should be abandoned? Is I/sigI a much better criterion than Rmerge? Lets look at the definitions: I is measured as a number of detector counts in the reflection minus background counts. sigI is measured as sq. root of I plus standard deviation (SD) for the background plus various deviations from ideal experiment (like noise from satellite crystals). Obviously, sigI cannot be measured accurately. Moreover, the 'resolution' is related to errors in the structural factors, which are average from several measurements. Errors in their scaling would affect the 'resolution', and I/sigI does not detect them, but Rmerge does! Rmerge = (I - I) / n*I where n is the number of measurements for the same structural factor (data redundancy). When n - infinity, Rmerge = sigI/I . From my experience, redundancy = 3-4 gives a very good agreement between Rmerge and sigI/I. If sigI/I is significantly lower than Rmerge, it means that the symmetry related reflections did not merge well. Under those conditions, Rmerge becomes a much better criterion for estimation of the 'resolution' than sigi/I. I AGREE THAT Rmerge=0.5 SHOULD NOT BE A CRITERION FOR DATA TRUNCATION. But we need a commonly accepted criterion to estimate the resolution, and Rmerge=0.5 is tested by the time. If someone decides to use I/sigI instead of Rmerge, fine, let it be 2.0. I do not know how it translates into CC... Alternatively, the resolution could be estimated from the electron density maps. But we need the commonly accepted rule how to do it, and it should be related to the old Rmerge=0.5 rule. I hope everyone agrees that the resolution should not be dead.. Alex On Jun 1, 2012, at 11:19 AM, Phil Evans wrote: As the K D paper points out, as the signal/noise declines at higher resolution, Rmerge goes up to infinity, so there is no sensible way to set a limiting value to determine resolution. That is not to say that Rmerge has no use: as you say it's a reasonably good metric to plot against image number to detect a problem. It just not a suitable metric for deciding resolution I/sigI is pretty good for this, even though the sigma estimates are not very reliable. CC1/2 is probably better since it is independent of sigmas and has defined values from 1.0 down to 0.0 as signal/noise decreases. But we should be careful of any dogma which says what data we should discard, and what the cutoff limits should be: I/sigI 3,2, or 1? CC1/2 0.2, 0.3, 0.5 ...? Usually it does not make a huge difference, but why discard useful data? Provided the data are properly weighted in refinement by weights incorporating observed sigmas (true in Refmac, not true in phenix.refine at present I believe), adding extra weak data should do no harm, at least out to some point. Program algorithms are improving in their treatment of weak data, but are by no means perfect. One problem as discussed earlier in this thread is that we have got used to the idea that nominal resolution is a single number indicating the quality of a structure, but this has never been true, irrespective of the cutoff method. Apart from the considerable problem of anisotropy, we all need to note the wisdom of Ethan Merritt We should also encourage people not to confuse the quality of the data with the quality of the model. Phil On 1 Jun 2012, at 18:59, aaleshin wrote: Please excuse my ignorance, but I cannot understand why Rmerge is unreliable for estimation of the resolution? I mean, from a theoretical point of view, 1/sigma is indeed a better criterion, but it is not obvious from a practical point of view. 1/sigma depends on a method for sigma estimation, and so same data processed by different programs may have different 1/sigma. Moreover, HKL2000 allows users to adjust sigmas manually. Rmerge estimates sigmas from differences between measurements of same structural factor, and hence
Re: [ccp4bb] @Ed: Death of Rmerge
I looked at those abstracts, thanks. Now I know that the K D paper means, promises to read it. Well, I agree that Rmerge should be replaced with Rrim (redundancy independent merging factor). Was not Z. Otwinowski first to use it in his scalepack? I agree, dependence on the number of measurements is such an obvious limitation of Rmerge that it should be corrected by the software developers long time ago. My point was that Rmerge (or its corrected version Rrim) better estimate the experimental errors than I/sigI, and so they are better suitable for estimation of the resolution. Also, Rmerge does not depend much on the number of measurements, when the data are highly redundant (and it is a common case nowadays). Alex On Jun 1, 2012, at 11:29 AM, Ed Pozharski wrote: http://www.nature.com/nsmb/journal/v4/n4/abs/nsb0497-269.html http://scripts.iucr.org/cgi-bin/paper?S0021889800018227 Just collect 360 sweep instead of 180 on a non-decaying crystal and see Rmerge go up due to increase in multiplicity (and enough with redundancy term - the extra data is not really *redundant*). Is your resolution worse or better? This has been argued over before. Rmerge has some value in comparing two datasets collected in perfectly identical conditions to see which crystal is better and it may predict to some extent what R-values you might expect. Otherwise, it's unreliable. Given that it's been 15 years since this was pointed out in no less than Nature group magazine, and we still hear that Rmerge should decide resolution cutoff, chances are increasingly slim that I will personally see the dethroning of that other major oppressor, R-value. On Fri, 2012-06-01 at 10:59 -0700, aaleshin wrote: Please excuse my ignorance, but I cannot understand why Rmerge is unreliable for estimation of the resolution? I mean, from a theoretical point of view, 1/sigma is indeed a better criterion, but it is not obvious from a practical point of view. 1/sigma depends on a method for sigma estimation, and so same data processed by different programs may have different 1/sigma. Moreover, HKL2000 allows users to adjust sigmas manually. Rmerge estimates sigmas from differences between measurements of same structural factor, and hence is independent of our preferences. But, it also has a very important ability to validate consistency of the merged data. If my crystal changed during the data collection, or something went wrong with the diffractometer, Rmerge will show it immediately, but 1/sigma will not. So, please explain why should we stop using Rmerge as a criterion of data resolution? Alex Sanford-Burnham Medical Research Institute 10901 North Torrey Pines Road La Jolla, California 92037 On Jun 1, 2012, at 5:07 AM, Ian Tickle wrote: On 1 June 2012 03:22, Edward A. Berry ber...@upstate.edu wrote: Leo will probably answer better than I can, but I would say I/SigI counts only the present reflection, so eliminating noise by anisotropic truncation should improve it, raising the average I/SigI in the last shell. We always include unmeasured reflections with I/sigma(I) = 0 in the calculation of the mean I/sigma(I) (i.e. we divide the sum of I/sigma(I) for measureds by the predicted total no of reflections incl unmeasureds), since for unmeasureds I is (almost) completely unknown and therefore sigma(I) is effectively infinite (or at least finite but large since you do have some idea of what range I must fall in). A shell with I/sigma(I) = 2 and 50% completeness clearly doesn't carry the same information content as one with the same I/sigma(I) and 100% complete; therefore IMO it's very misleading to quote I/sigma(I) including only the measured reflections. This also means we can use a single cut-off criterion (we use mean I/sigma(I) 1), and we don't need another arbitrary cut-off criterion for completeness. As many others seem to be doing now, we don't use Rmerge, Rpim etc as criteria to estimate resolution, they're just too unreliable - Rmerge is indeed dead and buried! Actually a mean value of I/sigma(I) of 2 is highly statistically significant, i.e. very unlikely to have arisen by chance variations, and the significance threshold for the mean must be much closer to 1 than to 2. Taking an average always increases the statistical significance, therefore it's not valid to compare an _average_ value of I/sigma(I) = 2 with a _single_ value of I/sigma(I) = 3 (taking 3 sigma as the threshold of statistical significance of an individual measurement): that's a case of comparing apples with pears. In other words in the outer shell you would need a lot of highly significant individual values 3 to attain an overall average of 2 since the majority of individual values will be 1. F/sigF is expected to be better than I/sigI because dx^2 = 2Xdx, dx^2/x^2 = 2dx/x, dI/I = 2* dF/F (or approaches that in the limit . . .) That depends
Re: [ccp4bb] @Phil-2:Death of Rmerge
Phil, I did not know issues that were discussed in this thread and so could not understand your explanation. Thanks to Dale Tronrud's email, which chewed it up for me, I now understand what was going on. I am totally with you on this matter. Actually, I was preaching same things, just calling them with different names: I am not a methods developer and my language is fool with working-class jargons... Alex On Jun 2, 2012, at 11:00 PM, aaleshin wrote: Could you please give me a reference to the K D paper? Without reading it, I do not see a problem with Rmerge going to infinity in high resolution shells. Indeed, I was taught at school that the crystallographic resolution is defined as a minimal distance between two peaks that can be distinguished in the electron density map. I was also taught that under normal conditions this would occur when the data are collected up to the shell, in which Rmerge = 0.5. One can collect more data (up to Rmerge=1.0 or even 100) but the resolution of the electron density map will not change significantly. I solved several structures of my own, and this simple rule worked every time. It failed only when the diffraction was very anisotropic, because the resolution was not uniform in all directions. But this obstacle can be easily overcome by presenting the resolution as a tensor with eigenvalues defined in the same simple rule (Rmerge = 0.5). Now, why such a simple method for estimation of the data resolution should be abandoned? Is I/sigI a much better criterion than Rmerge? Lets look at the definitions: I is measured as a number of detector counts in the reflection minus background counts. sigI is measured as sq. root of I plus standard deviation (SD) for the background plus various deviations from ideal experiment (like noise from satellite crystals). Obviously, sigI cannot be measured accurately. Moreover, the 'resolution' is related to errors in the structural factors, which are average from several measurements. Errors in their scaling would affect the 'resolution', and I/sigI does not detect them, but Rmerge does! Rmerge = (I - I) / n*I where n is the number of measurements for the same structural factor (data redundancy). When n - infinity, Rmerge = sigI/I . From my experience, redundancy = 3-4 gives a very good agreement between Rmerge and sigI/I. If sigI/I is significantly lower than Rmerge, it means that the symmetry related reflections did not merge well. Under those conditions, Rmerge becomes a much better criterion for estimation of the 'resolution' than sigi/I. I AGREE THAT Rmerge=0.5 SHOULD NOT BE A CRITERION FOR DATA TRUNCATION. But we need a commonly accepted criterion to estimate the resolution, and Rmerge=0.5 is tested by the time. If someone decides to use I/sigI instead of Rmerge, fine, let it be 2.0. I do not know how it translates into CC... Alternatively, the resolution could be estimated from the electron density maps. But we need the commonly accepted rule how to do it, and it should be related to the old Rmerge=0.5 rule. I hope everyone agrees that the resolution should not be dead.. Alex On Jun 1, 2012, at 11:19 AM, Phil Evans wrote: As the K D paper points out, as the signal/noise declines at higher resolution, Rmerge goes up to infinity, so there is no sensible way to set a limiting value to determine resolution. That is not to say that Rmerge has no use: as you say it's a reasonably good metric to plot against image number to detect a problem. It just not a suitable metric for deciding resolution I/sigI is pretty good for this, even though the sigma estimates are not very reliable. CC1/2 is probably better since it is independent of sigmas and has defined values from 1.0 down to 0.0 as signal/noise decreases. But we should be careful of any dogma which says what data we should discard, and what the cutoff limits should be: I/sigI 3,2, or 1? CC1/2 0.2, 0.3, 0.5 ...? Usually it does not make a huge difference, but why discard useful data? Provided the data are properly weighted in refinement by weights incorporating observed sigmas (true in Refmac, not true in phenix.refine at present I believe), adding extra weak data should do no harm, at least out to some point. Program algorithms are improving in their treatment of weak data, but are by no means perfect. One problem as discussed earlier in this thread is that we have got used to the idea that nominal resolution is a single number indicating the quality of a structure, but this has never been true, irrespective of the cutoff method. Apart from the considerable problem of anisotropy, we all need to note the wisdom of Ethan Merritt We should also encourage people not to confuse the quality of the data with the quality of the model. Phil On 1 Jun 2012, at 18:59, aaleshin wrote: Please excuse my ignorance, but I
Re: [ccp4bb] Death of Rmerge
Please excuse my ignorance, but I cannot understand why Rmerge is unreliable for estimation of the resolution? I mean, from a theoretical point of view, 1/sigma is indeed a better criterion, but it is not obvious from a practical point of view. 1/sigma depends on a method for sigma estimation, and so same data processed by different programs may have different 1/sigma. Moreover, HKL2000 allows users to adjust sigmas manually. Rmerge estimates sigmas from differences between measurements of same structural factor, and hence is independent of our preferences. But, it also has a very important ability to validate consistency of the merged data. If my crystal changed during the data collection, or something went wrong with the diffractometer, Rmerge will show it immediately, but 1/sigma will not. So, please explain why should we stop using Rmerge as a criterion of data resolution? Alex Sanford-Burnham Medical Research Institute 10901 North Torrey Pines Road La Jolla, California 92037 On Jun 1, 2012, at 5:07 AM, Ian Tickle wrote: On 1 June 2012 03:22, Edward A. Berry ber...@upstate.edu wrote: Leo will probably answer better than I can, but I would say I/SigI counts only the present reflection, so eliminating noise by anisotropic truncation should improve it, raising the average I/SigI in the last shell. We always include unmeasured reflections with I/sigma(I) = 0 in the calculation of the mean I/sigma(I) (i.e. we divide the sum of I/sigma(I) for measureds by the predicted total no of reflections incl unmeasureds), since for unmeasureds I is (almost) completely unknown and therefore sigma(I) is effectively infinite (or at least finite but large since you do have some idea of what range I must fall in). A shell with I/sigma(I) = 2 and 50% completeness clearly doesn't carry the same information content as one with the same I/sigma(I) and 100% complete; therefore IMO it's very misleading to quote I/sigma(I) including only the measured reflections. This also means we can use a single cut-off criterion (we use mean I/sigma(I) 1), and we don't need another arbitrary cut-off criterion for completeness. As many others seem to be doing now, we don't use Rmerge, Rpim etc as criteria to estimate resolution, they're just too unreliable - Rmerge is indeed dead and buried! Actually a mean value of I/sigma(I) of 2 is highly statistically significant, i.e. very unlikely to have arisen by chance variations, and the significance threshold for the mean must be much closer to 1 than to 2. Taking an average always increases the statistical significance, therefore it's not valid to compare an _average_ value of I/sigma(I) = 2 with a _single_ value of I/sigma(I) = 3 (taking 3 sigma as the threshold of statistical significance of an individual measurement): that's a case of comparing apples with pears. In other words in the outer shell you would need a lot of highly significant individual values 3 to attain an overall average of 2 since the majority of individual values will be 1. F/sigF is expected to be better than I/sigI because dx^2 = 2Xdx, dx^2/x^2 = 2dx/x, dI/I = 2* dF/F (or approaches that in the limit . . .) That depends on what you mean by 'better': every metric must be compared with a criterion appropriate to that metric. So if we are comparing I/sigma(I) with a criterion value = 3, then we must compare F/sigma(F) with criterion value = 6 ('in the limit' of zero I), in which case the comparison is no 'better' (in terms of information content) with I than with F: they are entirely equivalent. It's meaningless to compare F/sigma(F) with the criterion value appropriate to I/sigma(I): again that's comparing apples and pears! Cheers -- Ian
Re: [ccp4bb] Fwd: [ccp4bb] Death of Rmerge
There are things you can expect to learn from a 2Å structure that you are unlikely to learn from a 5Å structure, even if equal care has been given to both experiments, so it makes sense for the title to give the potential reader an idea which of the two cases is presented. But for this purpose it isn't going to matter whether 2Å is really 1.8Å or 2.2Å. What should the title say when a crystal diffracts to, lets say, 3 A in one direction and 4-5 A in others? Alex Aleshin, Sanford-Burnham Medical Research Institute 10901 North Torrey Pines Road La Jolla, California 92037 On May 31, 2012, at 2:50 PM, Ethan Merritt wrote: On Thursday, May 31, 2012 02:21:45 pm Dale Tronrud wrote: The resolution limit of the data set has been such an important indicator of the quality of the resulting model (rightly or wrongly) that it often is included in the title of the paper itself. Despite the fact that we now want to include more, weak, data than before we need to continue to have a quality indicator that readers can use to assess the models they are reading about. While cumbersome, one solution is to state what the resolution limit would have been had the old criteria been used, as was done in the paper you quote. This simply gives the reader a measure they can compare to their previous experiences. [\me dons flame suit] To the extent that reporting the resolution is simply a stand-in for reporting the quality of the model, we would do better to cut to the chase. For instance, if you map the Molprobity green/yellow/red model quality scoring onto good/mediocre/poor then you can title your paper Crystal Structure of Fabulous Protein Foo at Mediocre Quality [\me removes flame suit from back, and tongue from cheek] More seriously, I don't think it's entirely true that the resolution is reported as an indicator of quality in the sense that the model is well-refined. There are things you can expect to learn from a 2Å structure that you are unlikely to learn from a 5Å structure, even if equal care has been given to both experiments, so it makes sense for the title to give the potential reader an idea which of the two cases is presented. But for this purpose it isn't going to matter whether 2Å is really 1.8Å or 2.2Å. Now would be a good time to break with tradition and institute a new measure of quality of diffraction data sets. I believe several have been proposed over the years, but have simply not caught on. SFCHECK produces an optical resolution. Could this be used in the title of papers? I don't believe it is sensitive to the cutoff resolution and it produces values that are consistent with what the readers are used to. With this solution people could include whatever noisy data they want and not be guilty of overstating the quality of their model. We should also encourage people not to confuse the quality of the data with the quality of the model. Ethan -- Ethan A Merritt Biomolecular Structure Center, K-428 Health Sciences Bldg University of Washington, Seattle 98195-7742
Re: [ccp4bb] Akta vs HPLC
Back in Iowa State University we used Waters HPLC for protein purification during many years without noticeable damage to the stainless steel tubings. But Dan was right about the pumps, someone in the lab forgot to flush the high salt pump with water after its use and damaged the pump... Alex On May 29, 2012, at 5:14 PM, Daniel Anderson wrote: Hi, Ho, Your question has a lot of variables. HPLC columns should not be used on the Akta within my field of view because the Akta within my field of view does not have gradual pump acceleration and deceleration. HPLC columns can be damaged by sudden changes in pressure or composition. The HPLC within my field of view has wetted stainless steel surfaces and the mobile phase should not contain chloride ion or reductant. Chloride ion would accelerate corrosion of the stainless steel (and result in metal ions in the protein). Reductant would strip off the passivation (during maintenance I soak the stainless parts in nitric acid to keep them stainless) later resulting in corrosion. The Waters sales representative once told me that the pumps have to be salt-free and methanol-flushed at the end of every working day. Good luck implementing that policy. -Dan Ho Leung Ng wrote: Hello, My Akta Purifier is being repaired, and I'm thinking about borrowing a colleague's HPLC in the interim. What makes the Aktas different from HPLCs? I've used HPLCs for purifying small molecules and peptides but not proteins. Anything I should be careful about regarding keeping the machines, columns, and proteins happy? Thank you, Ho Ho Leung Ng University of Hawaii at Manoa Assistant Professor, Department of Chemistry h...@hawaii.edu mailto:h...@hawaii.edu
Re: [ccp4bb] Crystal behave funny
- Find a dinosaur from my generation who can suck one into a capillary and check diffraction at room T. - Try to find conditions where the crystals don't start to redissolve while you mount them As a matter of fact, people begin to forget that capillaries are good not only for checking the diffraction, but also for the data collection. However, the laws of Nature do not change with time, and the old dinosaur ways may still work. We just collected at SSRL a room T. data set from crystals that consisted by 75% v/v of water and 25% of a 2500 aa protein complex. The crystals did not like cryoprotectants (including Paratone), but were happy in the capillaries. On Apr 13, 2012, at 8:32 AM, Phoebe Rice wrote: I'd suggest: - Find a dinosaur from my generation who can suck one into a capillary and check diffraction at room T. - Try using those loops that look like miniature tennis paddles to give the crystal a little more support - To minimize strain on the crystal when pulling it out of the drop, try to get its long dimension perpendicular to the air-water interface (usually easier said than done). - Try to find conditions where the crystals don't start to redissolve while you mount them Original message Date: Fri, 13 Apr 2012 18:51:58 +0800 From: CCP4 bulletin board CCP4BB@JISCMAIL.AC.UK (on behalf of Prem kumar pk_ai...@yahoo.com) Subject: [ccp4bb] Crystal behave funny To: CCP4BB@JISCMAIL.AC.UK Hi all, I got some Protein + DNA complex crystals (image attached) recently. They are needle shape some times splitted chromosome type crystals. When we pick long needles they bend so much than normal crystal but they dont break. The small needle dissolve very fast as try to open the drop's film. we try to diffract the long needle crystals and they diffract up to 20 A resolution. Any suggestion how to improve those crystal packing. Thanks in advance! -Prem IMG_1438.JPG (2343k bytes)
Re: [ccp4bb] very informative - Trends in Data Fabrication
Thank you Phil, for clarification of my point, but it appears as cheating in a current situation, when an author has to fit a three dimensional statistics into a one-dimentional table. Moreover, many of journal reviewers may never worked with the low-resolution data and understand importance of every A^3 counts. It is not clear to me how to report the resolution of data when it is 3A in one direction, 3.5A in another and 5A in the third. Alex On Apr 9, 2012, at 4:51 AM, Phil Evans wrote: On 8 Apr 2012, at 21:18, aaleshin wrote: What I suggested with respect to the PDB data validation was adding some additional information that would allow to independently validate such parameters as the resolution and data quality (catching of model fabrications would be a byproduct of this process). Does the current system allow to overestimate those parameters? I believe so (but I might be wrong, correct me!). Periodically, people ask at ccp4bb how to determine the resolution of their data, but some idiots may decide to do it on their own and add 30% of noise to their structural factors. As James mentioned, one does not need to be extremely smart to do so, moreover, such an idiot would have less restraints than an educated crystallographer, because the idiot believes that nobody would notice his cheating. His moral principles are not corrupted, because he thinks that the model is correct and no harm is done. But the harm is still there, because people are forced to believe the model more than it deserves. The question is still open to me about what percentage of PDB structures overestimates data quality in terms of resolution. Is it possible to make it less dependent on the opinion of persons submitting the data? We all have so different opinions about everything... Regards, Alex Aleshin Using the weak high resolution data in a structure determination is not cheating. We should use data out to the point where there is no more significant and as long as it helps the structure determination and refinement, provided that we are using appropriate statistical treatment of the errors. We have become addicted to the idea that resolution is a single indicator of quality, and that is a gross over-simplification. Resolution tells us how many data were used, not their quality nor the quality of the model. Phil
Re: [ccp4bb] very informative - Trends in Data Fabrication
It is a wonderful server indeed, but its default setting cuts the resolution at 3 sigma (if I remember correctly). It is too stringent in my opinion. Also, it is not clear to me whether to submit all data to the highest resolution point, or the data that come from the server? But then again, the question remains at what sigma level to cut them? Aex On Apr 9, 2012, at 9:46 AM, David Schuller wrote: On 04/09/12 12:32, Boaz Shaanan wrote: How about such a footnote to Table 1: The resolution of data is 3A in the a direction, 3.5A in b direction and 5A in the c direction Wouldn't this do the trick? Usually there's a requirement for a table of statistics, including completeness and R in the outer shell. In the case of anisotropic data, what constitutes the outer shell? This is not a rhetorical question, I have some anisotropic data myself and will be facing these questions when it comes time to publish. This looks like a good place to plug the UCLA MBI Diffraction Anisotropy Server, which I found to be useful: http://services.mbi.ucla.edu/anisoscale/ Cheers, -- === All Things Serve the Beam === David J. Schuller modern man in a post-modern world MacCHESS, Cornell University schul...@cornell.edu
Re: [ccp4bb] very informative - Trends in Data Fabrication
Hi Pavel, Reporting the table that you suggested would create more red flags for the reviewers and readers than explaining how to understand the resolution of my data. We need more studies into this issue (correlation between the resolution of anisotropic data and model quality). And there should be a common rule how to report and interpret such data (IMHO). Regards, Alex On Apr 9, 2012, at 11:02 AM, Pavel Afonine wrote: Hi Alex, It is not clear to me how to report the resolution of data when it is 3A in one direction, 3.5A in another and 5A in the third. can't be easier I guess: just switch from characterizing data sets with one single number (which is suboptimal, at least, as Phil pointed out earlier) and show statistics by resolution instead. For example, R-factors, data completeness, Fobs shown in resolution bins are obviously much more informative metrics then a single number. If you want to be even more sophisticated, you can. See for example: A program to analyze the distributions of unmeasured reflections J. Appl. Cryst. (2011). 44, 865-872 L. Urzhumtseva and A. Urzhumtsev Pavel
Re: [ccp4bb] very informative - Trends in Data Fabrication
Since I was the person who started a public outcry to do something, I shell explain myself to my critics. Similarly to all of you, I do not care much about those few instances of structure fabrication. I might put too much emphases on them to initiate the discussion, but they are, indeed, only tiny blips on the ocean of science. But, could they be tips of a huge iceberg? That was my concern. I believe that an enormous competition in science that we experience nowadays makes many of us desperate, and desperation forces people to cheat. Is current validation system at PDB good enough to catch various aspects of data cheating? Is there a simple but efficient way to make it more difficult and, hence, less desirable? Good sportsmen (in terms of sport abilities) sometimes get caught with taking performance enhancers. I bet everyone would do it if the drug control did not exist. Many sportsmen would do it against their will, just because there was no other way to win. Do not you think a similar situation can develop in science? I suppose as social animals we like to think we can trust and be trusted Well, I suppose that these two antagonistic abilities of social animals (trust and cheating) developed in parallel as means to promote the evolution. In a very hierarchical society with no legal means to change a social status, cheating has been an important tool to contribute ones genes to a society. The socially unjust societies still exist and their members may have a slightly different view on morality of cheating than those from just societies. Moreover, ability to cheat often correlates with the intellect. Could not it be called cheating when someone is told to do something in one way, but he does it in his own way, because he believes it is more efficient? When a scientist feels that he is right about validity of his results, but they do not look good enough to be sold to validators, he is supposed to do more research. But he is out of time, why not to hide weak spots of the work if he knows that the major conclusions are RIGHT? Even if someone will redo the work later, they will be reproduced, right? In my opinion, this is the major motif for cheating in science. What I suggested with respect to the PDB data validation was adding some additional information that would allow to independently validate such parameters as the resolution and data quality (catching of model fabrications would be a byproduct of this process). Does the current system allow to overestimate those parameters? I believe so (but I might be wrong, correct me!). Periodically, people ask at ccp4bb how to determine the resolution of their data, but some idiots may decide to do it on their own and add 30% of noise to their structural factors. As James mentioned, one does not need to be extremely smart to do so, moreover, such an idiot would have less restraints than an educated crystallographer, because the idiot believes that nobody would notice his cheating. His moral principles are not corrupted, because he thinks that the model is correct and no harm is done. But the harm is still there, because people are forced to believe the model more than it deserves. The question is still open to me about what percentage of PDB structures overestimates data quality in terms of resolution. Is it possible to make it less dependent on the opinion of persons submitting the data? We all have so different opinions about everything... People invented laws to create conditions when they can trust each other. Sociopaths who do not follow the rules get caught and excluded from a society, which maintains the trust. But when the trust is abused, it quickly disappears. Many of those who wrote on the matter expressed a strong opinion that the system is not broken and we should continue trusting each other. Great! I do not mind the status quo. Regards, Alex Aleshin On Apr 8, 2012, at 8:48 AM, James Holton wrote: On 4/2/2012 6:03 AM, herman.schreu...@sanofi.com wrote: If James Holton had been involved, the fabrication would not have been discovered. Herman Uhh. Thanks. I think? Apologies for remaining uncharacteristically quiet. I have been keeping up with the discussion, but not sure how much difference one more vote would make on the various issues. Especially since most of this has come up before. I agree that fraud is sick and wrong. I think backing up your data is a good idea, etc. etc. However, I seem to have been declared a leading expert on fake data, so I suppose I ought to say something about that. Not quite sure I want to volunteer to be the Defense Against The Dark Arts Teacher (they always seem to end badly). But, here goes: I think the core of the fraud problem lies in our need for models, and I mean models in the general scientific sense not just PDB files. Fundamental to the practice of science is coming up with a model that explains the observations you
Re: [ccp4bb] very informative - Trends in Data Fabrication
Dear John, Thank you for a very informative letter about the IUCr activities towards archiving the experimental data. I feel that I did not explain myself properly. I do not object archiving the raw data, I just believe that current methodology of validating data at PDB is insufficiently robust and requires a modification. Implementation of the raw image storage and validation will take a considerable time, while the recent incidents of a presumable data frauds demonstrate that the issue is urgent. Moreover, presenting the calculated structural factors in place of the experimental data is not the only abuse that the current validation procedure encourages to do. There might be more numerous occurances of data massaging like overestimation of the resolution or data quality, the system does not allow to verify them. IUCr and PDB follows the American taxation policy, where the responsibility for a fraud is placed on people, and the agency does not take sufficient actions to prevent it. I believe it is inefficient and inhumane. Making a routine check of submitted data at a bit lower level would reduce a temptation to overestimate the unclearly defined quality statistics and make the model fabrication more difficult to accomplish. Many people do it unknowingly, and catching them afterwards makes no good. I suggested to turn the current incidence, which might be too complex for burning heretics, into something productive that is done as soon as possible, something that will prevent fraud from occurring. Since my persistent trolling at ccp4bb did not take any effect (until now), I wrote a bad-English letter to the PDB administration, encouraging them to take urgent actions. Those who are willing to count grammar mistakes in it can reading the message below. With best regards, Alexander Aleshin, staff scientist Sanford-Burnham Medical Research Institute 10901 North Torrey Pines Road La Jolla, California 92037 Dear PDB administrators; I am wringing to you regarding the recently publicized story about submission of calculated structural factors to the PDB entry 3k79 (http://journals.iucr.org/f/issues/2012/04/00/issconts.html). This presumable fraud (or a mistake) occurred just several years after another, more massive fabrication of PDB structures (Acta Cryst. (2010). D66, 115) that affected many scientists including myself. The repetitiveness of these events indicates that the current mechanism of structure validation by PDB is not sufficiently robust. Moreover, it is completely incapable of detecting smaller mischief such as overestimation of the data resolution and quality. There are two approaches to handling fraud problems: (1) raising policing and punishment, or (2) making a fraud too difficult to implement. Obviously, the second approach is more humane and efficient. This issue has been discussed on several occasions by the ccp4bb community, and some members began promoting the idea of submitting raw crystallographic images as a fraud repellent. However, this validation approach is not easy and cheap, moreover, it requires a considerable manpower to conduct it on a day-to-day basis. Indeed, indexing data sets is sometimes a nontrivial problem and cannot be accomplished automatically. For this reason, submitting the indexed and partially integrated data (such as .x files from HKL2000 or the output.mtz file from Mosfilm) appears as a cheaper substitute to the image storing/validating. Analysis of the partially integrated data provides almost same means to the fraud prevention as the images. Indeed, the observed cases of data fraud suggest that they would likely be attempted by a biochemist-crystallographer, who is insufficiently educated to fabricate the partially processed data. A method developer, on contrary, does not have a reasonable incentive to forge a particular structure, unless he teams up with a similarly minded biologist. But the latter scenario is very improbable and has not been detected yet. The most valuable benefit in using the partially processed data as a validation tool would be the standardization of definition for the data resolution and detection of inappropriate massaging of experimental data. Implementation of this approach requires minuscule adaptation of the current system, which most of practicing crystallographers would accept (in my humble opinion). The requirement to the data storage would be only ~1000 fold higher than the current one, and transferring the new data to PDB could be still done over the Internet. Moreover, storing the raw data is not required after the validation is done. A program such as Scala of CCP4 could be easily adopted to process the validation data and compare them with a conventional set of structural factors. Precise consistency of the two sets is not necessary. They only need to agree within statistically
Re: [ccp4bb] very informative - Trends in Data Fabrication
the trust we have in each others results. No one has time or resources to check everything, so science is based on trust. There are efforts underway outside crystallographic circles to address this larger threat to all science, and we should be participating in those discussions as much as possible. Ron On Thu, 5 Apr 2012, aaleshin wrote: Dear John,Thank you for a very informative letter about the IUCr activities towards archiving the experimental data. I feel that I did not explain myself properly. I do not object archiving the raw data, I just believe that current methodology of validating data at PDB is insufficiently robust and requires a modification. Implementation of the raw image storage and validation will take a considerable time, while the recent incidents of a presumable data frauds demonstrate that the issue is urgent. Moreover, presenting the calculated structural factors in place of the experimental data is not the only abuse that the current validation procedure encourages to do. There might be more numerous occurances of data massaging like overestimation of the resolution or data quality, the system does not allow to verify them. IUCr and PDB follows the American taxation policy, where the responsibility for a fraud is placed on people, and the agency does not take sufficient actions to prevent it. I believe it is inefficient and inhumane. Making a routine check of submitted data at a bit lower level would reduce a temptation to overestimate the unclearly defined quality statistics and make the model fabrication more difficult to accomplish. Many people do it unknowingly, and catching them afterwards makes no good. I suggested to turn the current incidence, which might be too complex for burning heretics, into something productive that is done as soon as possible, something that will prevent fraud from occurring. Since my persistent trolling at ccp4bb did not take any effect (until now), I wrote a bad-English letter to the PDB administration, encouraging them to take urgent actions. Those who are willing to count grammar mistakes in it can reading the message below. With best regards, Alexander Aleshin, staff scientist Sanford-Burnham Medical Research Institute 10901 North Torrey Pines Road La Jolla, California 92037 Dear PDB administrators; I am wringing to you regarding the recently publicized story about submission of calculated structural factors to the PDB entry 3k79 (http://journals.iucr.org/f/issues/2012/04/00/issconts.html). This presumable fraud (or a mistake) occurred just several years after another, more massive fabrication of PDB structures (Acta Cryst. (2010). D66, 115) that affected many scientists including myself. The repetitiveness of these events indicates that the current mechanism of structure validation by PDB is not sufficiently robust. Moreover, it is completely incapable of detecting smaller mischief such as overestimation of the data resolution and quality. There are two approaches to handling fraud problems: (1) raising policing and punishment, or (2) making a fraud too difficult to implement. Obviously, the second approach is more humane and efficient. This issue has been discussed on several occasions by the ccp4bb community, and some members began promoting the idea of submitting raw crystallographic images as a fraud repellent. However, this validation approach is not easy and cheap, moreover, it requires a considerable manpower to conduct it on a day-to-day basis. Indeed, indexing data sets is sometimes a nontrivial problem and cannot be accomplished automatically. For this reason, submitting the indexed and partially integrated data (such as .x files from HKL2000 or the output.mtz file from Mosfilm) appears as a cheaper substitute to the image storing/validating. Analysis of the partially integrated data provides almost same means to the fraud prevention as the images. Indeed, the observed cases of data fraud suggest that they would likely be attempted by a biochemist-crystallographer, who is insufficiently educated to fabricate the partially processed data. A method developer, on contrary, does not have a reasonable incentive to forge a particular structure, unless he teams up with a similarly minded biologist. But the latter scenario is very improbable and has not been detected yet. The most valuable benefit in using the partially processed data as a validation tool would be the standardization of definition for the data resolution and detection of inappropriate massaging of experimental data. Implementation of this approach requires minuscule adaptation of the current system, which most of practicing crystallographers would accept (in my humble opinion). The requirement to the data storage would be only ~1000 fold higher than the current one, and transferring the new data to PDB could be still
Re: [ccp4bb] very informative - Trends in Data Fabrication
Alright, if the image deposition is the only way out, then I am for it, but please make sure that synchrotrons will do it for me... On Apr 5, 2012, at 7:58 PM, Bernhard Rupp (Hofkristallrat a.D.) wrote: Ojweh c) Discarding your primary data is generally considered bad form... Agreed, but it is a big burden on labs to maintain archives of their raw data indefinitely. Even IRS allows to discard them after some time. But you DO have to file in the first place, right? How long to keep is an entirely different question. What is wrong with partially integrated data in terms of structure validation? Who thinks something is wrong with that idea? Section 3.1 under figure 3 of said incendiary pamphlet states: '...yadayadawhen unmerged data or images for proper reprocessing are not available owing to the unfortunate absence of a formal obligation to deposit unmerged intensity data or diffraction images.' They did not generate the bad data. This is a genuine American thinking! Ok, the US citizens on BB might take this one up on my behalf, gospodin ;-) видеть вас на Лубянке. But they might create conditions that would prevent their deposition. Sure. We are back to the 2007 Reid shoe bomber argument. If you make PDB deposition a total pain for everybody, you don't get compliance, you get defiance. Ever seen any happy faces in a TSA check line? Anyhow, image deposition will come. Over and out, BR
Re: [ccp4bb] very informative - Trends in Data Fabrication
Did you play as a child a game called a broken phone? It is when someone tells something quickly to a neighbor, and so on until the words come back to the author. Very funny game. My original thesis was that downloading/depositing the raw images would be a pain in the neck for crystallographers, so why would not to begin with the partially processed data, like .x files from HKL2000? People should be trained to hardships gradually... On Apr 5, 2012, at 8:57 PM, Bosch, Juergen wrote: How should they ? They have no clue which of the 20 datasets was actually useful to solve your structure. If you ask James Holton he has (suggested) to go back to the archived data after a certain time and try to solve the undeposited structures then :-) [Where is James anyhow ? Haven't seen a post recently from him] Seriously, I think it is in our own interest to submit the corresponding images which led to a structure solution somewhere. And as others mentioned bad data or good data can always serve for educational purposes. Just as an example http://strucbio.biologie.uni-konstanz.de/xdswiki/index.php/1Y13 Jürgen On Apr 5, 2012, at 11:46 PM, aaleshin wrote: Alright, if the image deposition is the only way out, then I am for it, but please make sure that synchrotrons will do it for me... On Apr 5, 2012, at 7:58 PM, Bernhard Rupp (Hofkristallrat a.D.) wrote: Ojweh c) Discarding your primary data is generally considered bad form... Agreed, but it is a big burden on labs to maintain archives of their raw data indefinitely. Even IRS allows to discard them after some time. But you DO have to file in the first place, right? How long to keep is an entirely different question. What is wrong with partially integrated data in terms of structure validation? Who thinks something is wrong with that idea? Section 3.1 under figure 3 of said incendiary pamphlet states: '...yadayadawhen unmerged data or images for proper reprocessing are not available owing to the unfortunate absence of a formal obligation to deposit unmerged intensity data or diffraction images.' They did not generate the bad data. This is a genuine American thinking! Ok, the US citizens on BB might take this one up on my behalf, gospodin ;-) видеть вас на Лубянке. But they might create conditions that would prevent their deposition. Sure. We are back to the 2007 Reid shoe bomber argument. If you make PDB deposition a total pain for everybody, you don't get compliance, you get defiance. Ever seen any happy faces in a TSA check line? Anyhow, image deposition will come. Over and out, BR .. Jürgen Bosch Johns Hopkins University Bloomberg School of Public Health Department of Biochemistry Molecular Biology Johns Hopkins Malaria Research Institute 615 North Wolfe Street, W8708 Baltimore, MD 21205 Office: +1-410-614-4742 Lab: +1-410-614-4894 Fax: +1-410-955-2926 http://web.mac.com/bosch_lab/
Re: [ccp4bb] very informative - Trends in Data Fabrication
People who raise their voices for a prolonged storage of raw images miss a simple fact that the volume of collected data increases proportionally if not faster than the cost of storage space drops. I just had an opportunity to collect data with the PILATUS detector at SSRL and say you that monster allows slicing the data 4-5 times thinner than other detectors do. Some people also like collecting very redundant data sets. Even now, transferring and storage of raw data from a synchrotron is a pain in the neck, but in a few years it may become simply impractical. And all this hassle is for the only real purpose of preventing data fraud? An't there a cheaper and more adequate solutions to the problem? I also wonder why after the first occurrence of data fraud several years ago, PDB did not take any action to prevent its appearance in the future? Or administrative actions are simply impossible nowadays without a mega-dollar grant? On Apr 4, 2012, at 3:45 PM, Eric Bennett wrote: Then everyone's data can be lost at once in the next cloud failure. Progress! The hardware failed in such a way that we could not forensically restore the data. What we were able to recover has been made available via a snapshot, although the data is in such a state that it may have little to no utility... -Amazon to some of its cloud customers following their major crash last year http://articles.businessinsider.com/2011-04-28/tech/29958976_1_amazon-customer-customers-data-data-loss -Eric On Apr 3, 2012, at 9:22 PM, Zhijie Li wrote: Hi, Regarding the online image file storage issue, I just googled cloud storage and had a look at the current pricing of such services. To my surprise, some companies are offering unlimited storage for as low as $5 a month. So that's $600 for 10 years. I am afraid that these companies will feel really sorry to learn that there are some monsters called crystallographers living on our planet. In our lab, some pre-21st century data sets were stored on tapes, newer ones on DVD discs and IDE hard drives. All these media have become or will become obsolete pretty soon. Not to mention the positive relationship of getting CRC errors with the medium's age. Admittedly, it may become quite a job to upload all image files that the whole crystallographic community generates per year. But for individual labs, I think clouding data might become something worth thinking of. Zhijie
Re: [ccp4bb] simple solution to - Trends in Data Fabrication
Hi James, My previous message on this matter remains unnoticed, but I also suggested a very simple solution to the data fraud: the crystallographers should submit to PDB partially processed data, like unmerged partial reflections. These files are much smaller than the images, and only a few people in the world are capable to forge them. This simple solution would kill any attempt to fabricate crystallographic data. Alex On Apr 3, 2012, at 7:11 AM, James Whisstock wrote: Hi I was thinking about the last statement in the Acta editorial - It is important to note, however, that in neither of these cases was a single frame of data collected. Not one.. This brought me back to the images.. To date there is no global acceptance that original diffractiom images must be deposited (though I personally think there should be). Many of the arguments around this issue relate to the time and space required to house such data. However (and apologies if this has already been raised and I have missed it), if our sole intent is to ascertain that there's no trouble at t'mill then deposition of a modest wedge of data and / or a 0 and 90, while not ideal, may be sufficient to provide a decent additional check and balance, particularly if such images, headers etc were automatically analysed as part of the already excellent validation tools in development. I'm sure there are a number of clever ways (that could be unadvertised or kept confidential to the pdb) that could be used to check off sufficient variables within such data such that it should (?) be very difficult to falsify images without triggering alarm bells. Of course this would probably then drive those that are truly bonkers to attempt to fabricate realistically noisy false diffraction images, however I would hope that such a scheme might make things just a little more difficult for those with fraudulent intent, particularly if no one (apart from the developers) knows precisely how and what the checking software checks! While it seems sad that it's come to this cell biologists and biochemists have had to deal with more and more sophisticated versions of the photoshopped western for years. Accordingly, most high profile journals run figures through commercial software that does a reasonable job of detection of such issues. J Sent from my iPhone On 03/04/2012, at 11:10 PM, Dyda d...@ulti.niddk.nih.gov wrote: I think that to review a paper containing a structure derived from crystallographic data should indeed involve the referee having access to coordinates and to the electron density. Without this access it is not possible to judge the quality and very often even the soundness of statements in the paper. I think the argument that this may give a competitive advantage to the referee who him or herself maybe working on the same thing should be mute, as I thought article refereeing was supposed to be a confidential process. Breaching this would be a serious ethical violation. In my experience, before agreeing to review, we see the abstract, I was always thought that I was supposed to decline if there is a potential conflict with my own work. Perhaps naively, but I always assumed that everyone acts like this. Unfortunately however, there is another serious issue. After a very troubling experience with a paper I reviewed, I discussed this issue with journal editors. What they said was that they already have a hell of time to find people who agree to referee, by raising the task level (asking refs to look at coords and density) they feared that no one would agree. Actually, perhaps many have noticed the large number of 5 liner referee reports saying really not much about a full length research article. People simply don't have the time to put the effort in. So I am not sure how realistic is to ask even more, for something that at some level, is pro bono work. Fred [32m*** Fred Dyda, Ph.D. Phone:301-402-4496 Laboratory of Molecular BiologyFax: 301-496-0201 DHHS/NIH/NIDDK e-mail:fred.d...@nih.gov Bldg. 5. Room 303 Bethesda, MD 20892-0560 URGENT message e-mail: 2022476...@mms.att.net Google maps coords: 39.000597, -77.102102 http://www2.niddk.nih.gov/NIDDKLabs/IntramuralFaculty/DydaFred ***[m
Re: [ccp4bb] Requested: Three-Day Data Fabrication Workshop
Dear James, With all due respect, you have left out a key component to successful data fabrication in the modern age: THE MOLECULAR REPLACEMENT. Since almost all new structures have more or less close homologues in PDB, a smart fabricator should use their experimental data as a template. It will be more difficult to detect than the data built from the calculated structural factors. To prevent future fabrication attempts, we do not need submitting detector images, partially processed structural data such as unmerged structural factors would work, and they do not take that much space. The switch to the new format could be done in no time Alex On Apr 2, 2012, at 8:39 AM, James Kiefer wrote: Dear Jacob, With all due respect, you have left out a key component to successful data fabrication in the modern age: software. It is quite obtuse not to have allocated at least one day of the workshop for practical applications of Photoshop to diffraction image generation and at least a passing coverage of whether or not Adobe Lightroom and crystallographic presets therein will be sufficiently capable of muddling the RCSB staff analysis of data feasibility checking. I would very much like to see Gerard Bricogne present a keynote lecture entitled something like, The R-Fake Parameter: A Maximum Likelihood Modulus to Define a Minimum Acceptable Data Drift Coefficient for Use in the Fabrication of Credibly Artificial Diffraction Data. I also believe that we are perhaps full of hubris as a crystallographic community, because an entire field of faked structural data has existed long before crystallographers even considered manufacturing their data. Specifically, the molecular modeling community has already surpassed us in their thinking on the subject. While we idly discuss how to properly generate false data, they have had the foresight to abandon ALL data...and even the starting coordinates in crystal structures - be they real or fictitious - and publish volumes of papers entirely unencumbered by reality or plausibility. My hat is off to them. Best regards, Jim On Mon, Apr 2, 2012 at 8:15 AM, Jacob Keller j-kell...@fsm.northwestern.edu wrote: Dear CCP4BB, due to increasing demand, it seems we should put together a workshop on data fabrication, covering the various important topics (chaired by JHo): --Images: the future of fabrication? How long can we rely on database Luddism? --Ways out: how to leave a trail of accidental data mix-ups --Publish large or small? Cost-benefit analyses of impact factor vs. risk of being discovered --Pushing the envelope: how significant is two [sic] significant --Crossing discipline boundaries: are data fabrication procedures universal? --Build a better hofkristallrat-trap: utilization of rhetorical bombast and indignation in reply letters --Break-out support-session with survivors: comforting words on careers after the fall --Session on the inextricably-related topic of grammatical pedantry, to be followed by a soccer (football?) match Greeks Vs. Latins Ample funding will be available from big pharma and other industry sectors Please submit further topics to the CCP4BB list JPK ps I can't believe no one mentioned the loathsome Latino-Greek multimer in the recent curmudgeonry postings. *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program email: j-kell...@northwestern.edu *** -- James Kiefer, Ph.D. Structural Biology Genentech, Inc. 1 DNA Way, Mailstop 27 South San Francisco, CA 94080-4990
Re: [ccp4bb] large R-Rfree difference in final structure
Careina, I recommend to compare the quality of your data (Rmerge) to that of an average data set of same resolution. Do you have meaningful data to 2.3A resolution? Another possibility is the anisotropicity of your data. Try this server http://services.mbi.ucla.edu/anisoscale/, if the anisotropicity is significant. Alex On Jul 13, 2011, at 8:38 AM, Careina Edgooms wrote: Dear ccp4 bulletin board I just have a slight concern regarding my Rwork Rfree difference. I have a structure that I have solved. I am reasonably content that it is complete because it has refined well, it no longer has bad geometries and contacts and all the rotamers, ramachandra, bond lengths etc are good. It gives favourable scores on molprobity and procheck. My only concern is the R factor difference. The resolution of the structure is 2.3A. The R factor is 0.24 after refinement but the Rfree is 0.33 which seems to me to be rather high. Should I be concerned? During refinement Rfree only drops from about 0.36 to 0.33 while the R factor drops from 0.31 to 0.24.. I have removed automatic weighting in refmac in order to constrain my bond lengths and angles during a couple of rounds of refinement. This did not have any effect on the R factors, however. I am fairly content that the space group I have chosen is correct so I am not sure what else could cause the big difference in R factors? There is no twinning. Can I be satisfied that my structure is correct despite the high R free or should I be doing other checks/ trying other things before I can submit this structure? Thank you for any help Careina
Re: [ccp4bb] Off topic - transformation problems
I never used pET20b, but I found on the Internet that it expresses inserts constitutively. Since the cloned protease should be toxic to cells, its constitutive expression might prevent the formation of colonies. But there might be millions of other reasons. Why did you use pET20b? http://www.biovisualtech.com/bvplasmid/pET-20b%28+%29.htm Alex On Jul 13, 2011, at 5:54 PM, Wonjin Bae wrote: Hi, all Recently, some bactrial source enzyme(sialidase) was subcloned pET20b and pGEX4T3 vector. Then, these two construct were transformed to BL21(DE3) expression host. DNA sequencing results were accurate. In case of pGEX, many colony was formed and shown the high level of expression. But, colony was not shown in pET20b case. (no one colony) In case of mock-pET20a vector transformed very well. What's the problems? I never experieced transfomation problems. Does anyone know how to solve the this problems? Thanks a lot !! Genie,
Re: [ccp4bb] Kd's in Crystals
Jacob, In case if the hint that I sent yesterday was not clear, below is the solution for the equation Kd=[P][L]/[PL] in terms of ligand occupancy: O=[ PL]/[Po]= 1/(Kd/L+1) You see, it does not depend on [Po] Alex On Jun 26, 2011, at 10:05 AM, aaleshin wrote: The concentration of a protein in a crystal [Po] and the volume of a crystal V are needed only to calculate the total amount of a ligand [Lo] required for soaking. [Lo] [Po]*V The occupancy of the active sites in a crystal will depend only on the ligand concentration in solution and Kd. It does not depend on protein concentration in the crystal. Indeed: Kd=[P][L]/[PL] Assuming total concentration of the protein = Po, Kd= 1mM and S= 1 mM, the active site occupancy will be: 1= P/Po-P; P/Po=1/2 So the concentration of the ligand in solution should be Kd to get the full occupancy. Alex
Re: [ccp4bb] Kd's in Crystals
Jacob, In the formula: Kd=[P][L]/[PL] [P] and [L] are concentrations of UNBOUND protein and ligand, and [PL] is that in the complex. Since the occupancy of the ligand in the crystal is [ PL]/[Po]= 1/(Kd/L+1), varying [L] around Kd like from 0.1Kd to 10Kd will make the titration of occupancy. You can calculate from the provided formula which [L] will give 0.25, 0.5 and 0.75 occupancies. Forget that the protein is crystallized. We assume that its behavior has not changed due to it. In reality, ligand affinity of conformationally flexible proteins can change by many orders of magnitude in both directions. This is why soaking does not work sometimes and you have to do co-crystallization. If you decide to titrate a crystal with a ligand, you should collect data and refine the ligandless and fully-ocupied crystals first, then use the superimposition of their structures for refinement of all other cases. Take care of waters that substitute for the partially bound ligand, they should have occupancies =1-Occ_of_ligand. Good luck. Alex On Jun 27, 2011, at 10:04 AM, Jacob Keller wrote: Yes, I think you are right--the somewhat counterintuitive case I was thinking of was, for example, when: Kd = 20nM [L] = 20uM [Po in crystal] = 20mM In this case, even though [L] = 20uM, since [L] is 1000 x Kd, the occupancy should be ~100%, and [PL] at equilibrium should be about 20mM, so in the crystal, the total [L] should be ~20mM. This explains, among other things, why bromophenol blue makes crystals bluer than the surrounding solution--the Kd is probably significantly lower than the BB concentration in the drop. Thanks for your clarifications! Jacob The question would remain, then, whether there is any utility in titrating ligands into crystals, and monitoring occupancies as a readout for binding. Although crystallization conditions are horribly non-physiological, perhaps there would be utility in the case where there are multiple known binding sites of various affinities, and other methods would have trouble resolving the binding events. One could start with: 1. totally saturated conditions, set occ=1 for all sites, refine B's, then 2. fix B's at this value, and refine the occ's in a subsequent series of dilutions. All of this is not totally theoretical--I am considering a set of experiments along these lines, where there really are multiple sites of varying affinity. *** Jacob Pearson Keller Northwestern University Medical Scientist Training Program cel: 773.608.9185 email: j-kell...@northwestern.edu ***
Re: [ccp4bb] Kd's in Crystals
The concentration of a protein in a crystal [Po] and the volume of a crystal V are needed only to calculate the total amount of a ligand [Lo] required for soaking. [Lo] [Po]*V The occupancy of the active sites in a crystal will depend only on the ligand concentration in solution and Kd. It does not depend on protein concentration in the crystal. Indeed: Kd=[P][L] / [PL] (chemical equilibrium equation) Assuming total concentration of the protein = Po, Kd= 1mM and S= 1 mM, the active site occupancy will be: 1= P/Po-P; P/Po=1/2 So the concentration of the ligand in solution should be Kd to get the full occupancy. Alex On Jun 25, 2011, at 9:19 PM, Jacob Keller wrote: Upon some reflection, I think one can say this: first, let's say the protein in question is 30kD, with a solvent content of 50%, and we know that solid protein density is ~1200mg/mL. Therefore, the protein concentration in the crystal would be ~20mM. Because Kd's assume infinitesimal ligand concentration, I think that neglecting ligand depletion effects mentioned by Edward Berry, say by having a huge reservoir or transferring the crystal to an appropriate soaking environment, that all ligands which bind with a better than ~20mM Kd should be bound in that crystal, even at extremely low ligand concentrations, so changing [ligand] from 1pM to 10mM should not change occupancy much, again assuming equilibrium and neglecting ligand depletion. JPK
Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford.
I also like our Nanodrop, but I do not recommend using it for Bradford measurements. The 25% accuracy mentioned by Flip is pretty good for biological samples. Using 50 ul cuvette in a traditional spectrophotometer will not give this accuracy because cleanness of the cuvette will be a big issue... Alex On Jun 16, 2011, at 12:43 PM, Oganesyan, Vaheh wrote: I completely disagree with Filip’s assessment. I’ve been using nanodrop nearly 5 years and never had inconsistency issues. If you work at reasonable speed (if you put a drop there then lower the lever and click measure before you do anything else) there will be no issues. At very high concentrations the accuracy and therefore consistency may become lower. Concentrations between 5 and 10 mg/ml should be fine. The instrument is pricey though. Vaheh From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Filip Van Petegem Sent: Thursday, June 16, 2011 3:34 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford. Dear Arnon, the Bradford method is not recommended for accurate measurements. The readings are strongly dependent on the amino acid composition. A much better method is using the absorption at 280nm under denaturing conditions (6M Guanidine), and using calculated extinction coefficients based on the composition of mostly Tyrosine and Tryptophan residues (+ disulfide bonds). This method is also old (Edelhoch, 1967), but very reliable. One thing about the nanodrop: smaller volume = more evaporation. On the demo we've had, I was so unimpressed with the precision (25% variability between two consecutive measurement) that we didn't consider this instrument at all. So unless you just want a 'rough' estimate, I wouldn't recommend it at all. But most respectable spectrophotometers will take cuvettes with 50ul volumes - a big step up from 1ml volumes... Filip Van Petegem On Thu, Jun 16, 2011 at 12:15 PM, Arnon Lavie la...@uic.edu wrote: Dear fellow crystallographers - a question about spectrophotometers for protein concentration determination. We are so last millennium - using Bradford reagent/ 1 ml cuvette for protein conc. determination. We have been considering buying a Nanodrop machine (small volume, no dilution needed, fast, easy). However, while testing our samples using a colleague's machine, we have gotten readings up to 100% different to our Bradford assay (all fully purified proteins). For example, Bradford says 6 mg/ml, Nanodrop 3 mg/ml. So while it is fun/easy to use the Nanodrop, I am not sure how reliable are the measurements (your thoughts?). So QUESTION 1: What are people's experience regarding the correlation between Nanodrop and Bradford? While researching the Nanodrop machine, I heard about the Implen NanoPhotmeter Pearl. So Question 2: Is the Pearl better/worse/same as the Nanodrop for our purpose? Thank you for helping us to advance to the next millennium, even if it is nearly a dozen years late. Arnon -- *** Arnon Lavie, Professor Dept. of Biochemistry and Molecular Genetics University of Illinois at Chicago 900 S. Ashland Ave. Molecular Biology Research Building, Room 1108 (M/C 669) Chicago, IL 60607 U.S.A. Tel:(312) 355-5029 Fax:(312) 355-4535 E-mail: la...@uic.edu http://www.uic.edu/labs/lavie/ *** -- Filip Van Petegem, PhD Assistant Professor The University of British Columbia Dept. of Biochemistry and Molecular Biology 2350 Health Sciences Mall - Rm 2.356 Vancouver, V6T 1Z3 phone: +1 604 827 4267 email: filip.vanpete...@gmail.com http://crg.ubc.ca/VanPetegem/ To the extent this electronic communication or any of its attachments contain information that is not in the public domain, such information is considered by MedImmune to be confidential and proprietary. This communication is expected to be read and/or used only by the individual(s) for whom it is intended. If you have received this electronic communication in error, please reply to the sender advising of the error in transmission and delete the original message and any accompanying documents from your system immediately, without copying, reviewing or otherwise using them for any purpose. Thank you for your cooperation.
Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford.
Filip, 25% accuracy is observed only for very diluted (OD280 0.1) or concentrated samples. But those sample a rarely used for ITC or CD. The concentrated samples require dilution but a regular spec does it too. Since the light passway is very short in Nanodrop it is accurate with more concentrated samples, which we crystallographers use, so Nanodrop is ideal instrument for our trade. If the drop is within recommended volume like 1-2 ul for our model, its size has a very small influence on the measurement. Cuvettes will give a better accuracy provided you clean them properly. I hated those times when I had to measure a concentration because of a need to wash a cuvette. In a biological lab they are always dirty. We switched to plastic disposable cuvettes for that reason... Alex On Jun 16, 2011, at 1:06 PM, Filip Van Petegem wrote: 25% is not acceptable for ITC or CD experiments though... I was just sharing our bad experience with a demo nanodrop we had. Even if evaporation is not an issue, one has to take pipetting errors into account when dealing with small volumes. The relative error on 1-2ul is a lot bigger than on 50ul. Unless you want to pre-mix 50ul and use a small quantity of that, which defeats the purpose of miniaturization... It all depends on your applications and sample availability, but if you want a very accurate measurement, miniaturized volumes just won't get you the same accuracy. Cuvettes will give a better accuracy provided you clean them properly. Just some water or EtOH is *not* enough... Filip Van Petegem On Thu, Jun 16, 2011 at 12:52 PM, aaleshin aales...@burnham.org wrote: I also like our Nanodrop, but I do not recommend using it for Bradford measurements. The 25% accuracy mentioned by Flip is pretty good for biological samples. Using 50 ul cuvette in a traditional spectrophotometer will not give this accuracy because cleanness of the cuvette will be a big issue... Alex On Jun 16, 2011, at 12:43 PM, Oganesyan, Vaheh wrote: I completely disagree with Filip’s assessment. I’ve been using nanodrop nearly 5 years and never had inconsistency issues. If you work at reasonable speed (if you put a drop there then lower the lever and click measure before you do anything else) there will be no issues. At very high concentrations the accuracy and therefore consistency may become lower. Concentrations between 5 and 10 mg/ml should be fine. The instrument is pricey though. Vaheh From: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] On Behalf Of Filip Van Petegem Sent: Thursday, June 16, 2011 3:34 PM To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford. Dear Arnon, the Bradford method is not recommended for accurate measurements. The readings are strongly dependent on the amino acid composition. A much better method is using the absorption at 280nm under denaturing conditions (6M Guanidine), and using calculated extinction coefficients based on the composition of mostly Tyrosine and Tryptophan residues (+ disulfide bonds). This method is also old (Edelhoch, 1967), but very reliable. One thing about the nanodrop: smaller volume = more evaporation. On the demo we've had, I was so unimpressed with the precision (25% variability between two consecutive measurement) that we didn't consider this instrument at all. So unless you just want a 'rough' estimate, I wouldn't recommend it at all. But most respectable spectrophotometers will take cuvettes with 50ul volumes - a big step up from 1ml volumes... Filip Van Petegem On Thu, Jun 16, 2011 at 12:15 PM, Arnon Lavie la...@uic.edu wrote: Dear fellow crystallographers - a question about spectrophotometers for protein concentration determination. We are so last millennium - using Bradford reagent/ 1 ml cuvette for protein conc. determination. We have been considering buying a Nanodrop machine (small volume, no dilution needed, fast, easy). However, while testing our samples using a colleague's machine, we have gotten readings up to 100% different to our Bradford assay (all fully purified proteins). For example, Bradford says 6 mg/ml, Nanodrop 3 mg/ml. So while it is fun/easy to use the Nanodrop, I am not sure how reliable are the measurements (your thoughts?). So QUESTION 1: What are people's experience regarding the correlation between Nanodrop and Bradford? While researching the Nanodrop machine, I heard about the Implen NanoPhotmeter Pearl. So Question 2: Is the Pearl better/worse/same as the Nanodrop for our purpose? Thank you for helping us to advance to the next millennium, even if it is nearly a dozen years late. Arnon -- *** Arnon Lavie, Professor Dept. of Biochemistry and Molecular Genetics University of Illinois at Chicago 900 S
Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford.
Mischa, You intrigued me. What is the experimental technique for the Extinction Coefficient measurement (which requires knowledge of protein concentration)? Let me guess, Bradford? Protein evaporation and weighing? Alex On Jun 16, 2011, at 4:22 PM, Machius, Mischa Christian wrote: With respect to the Edelhoch method and the ProtParam server, I would strongly recommend determining extinction coefficients experimentally and not rely on the ProtParam values. The reason is that the underlying extinction coefficients in the formula used by ProtParam and referenced there are statistical averages. They may or may not be valid for a given protein. I have seen differences of more than 20% between the theoretical and experimental extinction coefficients, particularly for proteins with few Trp and Tyr residues. When relying on relative concentrations, this inaccuracy is not detrimental, but when absolute concentrations are needed (CD, AUC, ITC, any binding experiment, etc.), such a difference would be considered huge. Determining an extinction coefficient experimentally takes but a few minutes. Cheers! MM On Jun 16, 2011, at 6:22 PM, Petr Leiman wrote: Totally support the statements below. We have had several proteins with A280 absorbance of 0.1 or less (at 1 mg/ml). You _have_ to use Bradford in the Nanodrop or whatnot to measure the concentration. Before purchasing the Nanodrop we used a Hellma TrayCell and a normal UV/Vis instrument. Similar to the Nanodrop, the sample volume in TrayCell is 2-3 ul. Traycell works well at a fraction of the Nanodrop cost, but Nanodrop is a lot more convenient to use for high concentration quick measurements (especially if you need to measure several things in succession), so you get what you pay for. Petr P.S. Expasy's Protparam tool has been around for ages (10-12+ years?). That plus the Nanodrop are two essential and synergetic tools of a protein chemist/crystallographer. On Jun 16, 2011, at 10:31 PM, Edward A. Berry wrote: Bradford is an assay, Nanodrop is a spectrophotometer. Both the A280 and Bradford methods are strongly dependent on amino acid composition, so unless you correct A280 for that as mentioned by Filip, either one is semiquantitative. Occasionally you come across a protein with no tryptophan which will have a much lower extinction coefficient. Try making a 1 g/l solution of gelatin (collagen?) and see what its A280 is! I noticed recently the protparam tool at http://ca.expasy.org/cgi-bin/protparam estimates the extinction coefficient given a sequence. David Briggs wrote: ~~~ I wouldn't touch Bradford with a barge-pole. I've found it to be wildly inaccurate for certain proteins I've handled, where as the OD280 measurements have been fine. One wonders what does fine mean, like same as with Biuret or Kjeldahl nitrogen, or solution made up by weight? --- Mischa Machius, PhD Director, Center for Structural Biology Assoc. Professor, Dept. of Pharmacology Member, Lineberger Comprehensive Cancer Center University of North Carolina 4079 Genetic Medicine CB#7365 120 Mason Farm Road Chapel Hill, NC 27599-7365, U.S.A. tel: +1-919-843-4485 fax: +1-919-966-5640 email: mach...@unc.edu
Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford.
Sorry for misprint, I meant evaporating water from a protein solution... On Jun 16, 2011, at 4:45 PM, aaleshin wrote: Mischa, You intrigued me. What is the experimental technique for the Extinction Coefficient measurement (which requires knowledge of protein concentration)? Let me guess, Bradford? Protein evaporation and weighing? Alex On Jun 16, 2011, at 4:22 PM, Machius, Mischa Christian wrote: With respect to the Edelhoch method and the ProtParam server, I would strongly recommend determining extinction coefficients experimentally and not rely on the ProtParam values. The reason is that the underlying extinction coefficients in the formula used by ProtParam and referenced there are statistical averages. They may or may not be valid for a given protein. I have seen differences of more than 20% between the theoretical and experimental extinction coefficients, particularly for proteins with few Trp and Tyr residues. When relying on relative concentrations, this inaccuracy is not detrimental, but when absolute concentrations are needed (CD, AUC, ITC, any binding experiment, etc.), such a difference would be considered huge. Determining an extinction coefficient experimentally takes but a few minutes. Cheers! MM On Jun 16, 2011, at 6:22 PM, Petr Leiman wrote: Totally support the statements below. We have had several proteins with A280 absorbance of 0.1 or less (at 1 mg/ml). You _have_ to use Bradford in the Nanodrop or whatnot to measure the concentration. Before purchasing the Nanodrop we used a Hellma TrayCell and a normal UV/Vis instrument. Similar to the Nanodrop, the sample volume in TrayCell is 2-3 ul. Traycell works well at a fraction of the Nanodrop cost, but Nanodrop is a lot more convenient to use for high concentration quick measurements (especially if you need to measure several things in succession), so you get what you pay for. Petr P.S. Expasy's Protparam tool has been around for ages (10-12+ years?). That plus the Nanodrop are two essential and synergetic tools of a protein chemist/crystallographer. On Jun 16, 2011, at 10:31 PM, Edward A. Berry wrote: Bradford is an assay, Nanodrop is a spectrophotometer. Both the A280 and Bradford methods are strongly dependent on amino acid composition, so unless you correct A280 for that as mentioned by Filip, either one is semiquantitative. Occasionally you come across a protein with no tryptophan which will have a much lower extinction coefficient. Try making a 1 g/l solution of gelatin (collagen?) and see what its A280 is! I noticed recently the protparam tool at http://ca.expasy.org/cgi-bin/protparam estimates the extinction coefficient given a sequence. David Briggs wrote: ~~~ I wouldn't touch Bradford with a barge-pole. I've found it to be wildly inaccurate for certain proteins I've handled, where as the OD280 measurements have been fine. One wonders what does fine mean, like same as with Biuret or Kjeldahl nitrogen, or solution made up by weight? --- Mischa Machius, PhD Director, Center for Structural Biology Assoc. Professor, Dept. of Pharmacology Member, Lineberger Comprehensive Cancer Center University of North Carolina 4079 Genetic Medicine CB#7365 120 Mason Farm Road Chapel Hill, NC 27599-7365, U.S.A. tel: +1-919-843-4485 fax: +1-919-966-5640 email: mach...@unc.edu
Re: [ccp4bb] Nanodrop versus Nanophotomter Pearl versus good old Bradford.
I see, by the experimental determination of the extinction coefficient you mean correction for the difference between unfolded (which can be computed accurately) and folded proteins. Am I right? Sorry for making this topic viral... Alex On Jun 16, 2011, at 5:06 PM, Machius, Mischa Christian wrote: The method is that by Edelhoch, mentioned a couple of times already in this discussion. It's also described in the paper by Pace et al., the same paper that the formula in ProtParam is from (ProtParam does not use the values determined by Gill von Hippel). Last time I looked into this, the consensus was that the Edelhoch method is the most accurate method for protein concentration determination; more accurate than dry-weighing plus N-terminal sequencing, etc. MM On Jun 16, 2011, at 7:51 PM, aaleshin wrote: Sorry for misprint, I meant evaporating water from a protein solution... On Jun 16, 2011, at 4:45 PM, aaleshin wrote: Mischa, You intrigued me. What is the experimental technique for the Extinction Coefficient measurement (which requires knowledge of protein concentration)? Let me guess, Bradford? Protein evaporation and weighing? Alex On Jun 16, 2011, at 4:22 PM, Machius, Mischa Christian wrote: With respect to the Edelhoch method and the ProtParam server, I would strongly recommend determining extinction coefficients experimentally and not rely on the ProtParam values. The reason is that the underlying extinction coefficients in the formula used by ProtParam and referenced there are statistical averages. They may or may not be valid for a given protein. I have seen differences of more than 20% between the theoretical and experimental extinction coefficients, particularly for proteins with few Trp and Tyr residues. When relying on relative concentrations, this inaccuracy is not detrimental, but when absolute concentrations are needed (CD, AUC, ITC, any binding experiment, etc.), such a difference would be considered huge. Determining an extinction coefficient experimentally takes but a few minutes. Cheers! MM On Jun 16, 2011, at 6:22 PM, Petr Leiman wrote: Totally support the statements below. We have had several proteins with A280 absorbance of 0.1 or less (at 1 mg/ml). You _have_ to use Bradford in the Nanodrop or whatnot to measure the concentration. Before purchasing the Nanodrop we used a Hellma TrayCell and a normal UV/Vis instrument. Similar to the Nanodrop, the sample volume in TrayCell is 2-3 ul. Traycell works well at a fraction of the Nanodrop cost, but Nanodrop is a lot more convenient to use for high concentration quick measurements (especially if you need to measure several things in succession), so you get what you pay for. Petr P.S. Expasy's Protparam tool has been around for ages (10-12+ years?). That plus the Nanodrop are two essential and synergetic tools of a protein chemist/crystallographer. On Jun 16, 2011, at 10:31 PM, Edward A. Berry wrote: Bradford is an assay, Nanodrop is a spectrophotometer. Both the A280 and Bradford methods are strongly dependent on amino acid composition, so unless you correct A280 for that as mentioned by Filip, either one is semiquantitative. Occasionally you come across a protein with no tryptophan which will have a much lower extinction coefficient. Try making a 1 g/l solution of gelatin (collagen?) and see what its A280 is! I noticed recently the protparam tool at http://ca.expasy.org/cgi-bin/protparam estimates the extinction coefficient given a sequence. David Briggs wrote: ~~~ I wouldn't touch Bradford with a barge-pole. I've found it to be wildly inaccurate for certain proteins I've handled, where as the OD280 measurements have been fine. One wonders what does fine mean, like same as with Biuret or Kjeldahl nitrogen, or solution made up by weight? --- Mischa Machius, PhD Director, Center for Structural Biology Assoc. Professor, Dept. of Pharmacology Member, Lineberger Comprehensive Cancer Center University of North Carolina 4079 Genetic Medicine CB#7365 120 Mason Farm Road Chapel Hill, NC 27599-7365, U.S.A. tel: +1-919-843-4485 fax: +1-919-966-5640 email: mach...@unc.edu --- Mischa Machius, PhD Director, Center for Structural Biology Assoc. Professor, Dept. of Pharmacology Member, Lineberger Comprehensive Cancer Center University of North Carolina 4079 Genetic Medicine CB#7365 120 Mason Farm Road Chapel Hill, NC 27599-7365, U.S.A. tel: +1-919-843-4485 fax: +1-919-966-5640 email: mach...@unc.edu
Re: [ccp4bb] 96-well block storage
I'm surprised that a better re-sealing system has not been invented to prevent evaporation from the blocks when they are stored. Companies that produce these screen want us to buy their screens as often as possible... We use transparent plastic seals from Hampton Research. The aluminum foil appears to seal better, but when we peel it off, the sticky layer may remain on top of wells and plug needles of a dispenser (we use Phoenix). Covering a block with Saran Wrap in addition to the seal also reduces the evaporation. Finally, storing a screen block in a refrigerator also helps, but this often causes precipitation of some solutions. I would also like to learn about a better storage technique. The other problem that we have here is mold (fungi) growing in the screen solutions. Our entire building is contaminated with mold and the PEG screens often get fungi growing in them. Alex On Jun 14, 2011, at 2:18 PM, Brian Mark wrote: Hi all, Considering the popularity of 96-well deep well block format for purchasing and storing protein crystallization conditions, I'm surprised that a better re-sealing system has not been invented to prevent evaporation from the blocks when they are stored. We typically don't consume our blocks fast enough to avoid this issue. There must be a better way to re-seal these blocks other than using peel and stick foil tape over and over again... Would anyone like to share their optimal method for storing 96-wells blocks that avoids (or a least minimizes) evaporation? Cheers, Brian Mark
Re: [ccp4bb] Using chaperones to boost expression in E. coli
Hi Michael, It worked for me with one viral enzyme but did not work with 2 others. In the successful case, I used Takara's chaperone plasmids to screen for the best composition of chaperones. The GroEL-GroES improved the yield of the soluble enzyme, but I got same activity (per 1L of media) as before. The plasmid with chaperone tig gave the biggest increase in activity (~4-8 fold). Still, the yield was rather low, in a range of 1 mg of purified protein per 1L of media. I also had to optimize the expression time and temperature to get best results. They were ~14 h at 17-20C. Interestingly, the protein failed to express in insect cells (was too poisonous for them), so the combination of a solubilizing domain, chaperones and Ecoli expression were the only available option for me. Regards Alex On Apr 22, 2011, at 8:28 AM, Kothe, Michael wrote: Dear ccp4bb, I am curious to hear of examples where the expression of well-behaved protein was achieved by the coexpression of chaperones in E. coli. I know the appropriate strains and vectors exist, but I can’t remember hearing of a successful case. I have heard anecdotally of several cases where it was tried without success (including one attempt I made myself). I also heard of concerns that the chaperones might copurify with the (now soluble) protein of interest and are difficult to remove. Thanks, Michael
Re: [ccp4bb] Reverse Translatase
Doesn't natural selection act like a Reverse Translatase? Which is quite an elegant implementation of the idea... On Sep 7, 2010, at 6:29 PM, Artem Evdokimov wrote: Regardless of whether a system like this exists in Nature or not - it's fun to imagine! On a microscopic scale one could propose a hypothetical mechanism by which a completely unfolded polypeptide chain could be fed into a gated (or state-locked) peptidase that may break the chain down in a co-ordinated stepwise fashion; releasing individual aa's into some sort of a nanoscale channel. The released aa's would then be sequentially coupled to something resembling tRNA - with pre-formed trinucleotides attached on the other end. Coupling would then presumably permit the triplets to ligate to one another sequentially - the resulting ssDNA or ssRNA would then have to be converted into a stable ds-form via the usual means, or otherwise protected in one of the usual ways. Codon space could be expanded by pre-loading carrier molecules with more than one type of triplet per carrier (biased towards whatever codon frequencies are prominent in the organism of choice) although this in no way resolves the random nature of the actual codon use within the resulting nucleotide sequence. The issue of amino acid coupling selectivity is pretty hairy - the best I could think of on a short notice is to have the receptor sites for individual aa's arranged in order of dropping selectivity -- however there is still the matter of shape/property similarities throwing wrenches into the works. An alternative would be a series of binary gates working on an exclusion principle. As to practicality of this kind of stuff - I am not sure; I can imagine an application similar to nano-scale multiparallel pyrosequencing: an unknown protein would be broken down into peptides via nonselective protease of some sort and then relatively short individual peptides are 'sequenced' in parallel, producing short DNA sequences that would later be complemented to dsDNA and allowed to cross-anneal and self-assemble via overlaps, similar to gapped gene assembly from short synthetic fragments (that first protease better be *really* non-specific!). At the end one could sequence the resulting long DNA to see what the original protein was like. A. On Tue, Sep 7, 2010 at 8:35 AM, David Schuller dj...@cornell.edu wrote: On 09/06/10 21:36, Jacob Keller wrote: Dear Crystallographers, does anyone know of any conceptual reason why a reverse translatase enzyme (protein--nucleic acid) could not exist? I can think of so many things for which such an enzyme would be helpful, both to cells and to scientists...! Unless there is something I am missing, it would seem to me conceptually almost impossible that it *not* exist. See: The RNA/Protein Symmetry Hypothesis: Experimental Support for Reverse Translation of Primitive Proteins Masayuki Nahimoto, J. Theor. Biol. (2001) 209, pp 181-187. In which Nahimoto proposes such a system, and additionally proposes that it actually existed early in the development of life on this planet. Reasons why it could not exist - No. Reasons why it would be very difficult - yes. And plenty of reasons why Nahimoto is probably wrong about it having actually existed: There is absolutely no evidence presented that such a system was ever in operation in the history of life on this planet. Current theories such as the RNA World are much more likely explanations for how life as we currently know it may have developed from a pre- biotic state. DNA replication, DNA=RNA transcription, and RNA=Protein translation all depend on nucleic acid base pairing for part of their specificity. It truly is the secret of life. And it would not be especially helpful in Protein=RNA reverse translation. Forward translation takes place in the ribosome, but extra specificity is smuggled in via a large set of tRNAs and tRNA charging enzymes, in reactions which took place beforehand, which are then made use of through the base-pairing codon:anti-codon recognition. Reverse translation would most definitely not be running forward translation in reverse; the specificity cannot be handled ahead of time, it needs to be available at the site of reverse translation itself when each successive peptide residue is presented. Progressivity: If different recognition sites are swapped in, this has to be done while keeping place in both the protein chain and in the growing nucleotide chain. Possibly the protein chain might be cleaved during the process. The chemistry and geometry of peptide residues is far more variable than that of nucleotide residues. The genetic code of reverse translation would be completely independent of that in forward translation. For the two to have matched up (in the proposed naturally occurring RT system) would have been extremely fortuitous, imposing a strong barrier to the introduction of such a system. Difficulty
Re: [ccp4bb] Reverse Translatase
I shell correct myself. The Darwin evolution of species is not sufficient to perform all functions of the reverse translatase. The Nature also uses viruses in order to translate proteins from different species. The other forms of reverse translation were probably not needed before the introduction of the immune system. Alex On Sep 7, 2010, at 7:26 PM, aaleshin wrote: Doesn't natural selection act like a Reverse Translatase? Which is quite an elegant implementation of the idea... On Sep 7, 2010, at 6:29 PM, Artem Evdokimov wrote: Regardless of whether a system like this exists in Nature or not - it's fun to imagine! On a microscopic scale one could propose a hypothetical mechanism by which a completely unfolded polypeptide chain could be fed into a gated (or state-locked) peptidase that may break the chain down in a co-ordinated stepwise fashion; releasing individual aa's into some sort of a nanoscale channel. The released aa's would then be sequentially coupled to something resembling tRNA - with pre-formed trinucleotides attached on the other end. Coupling would then presumably permit the triplets to ligate to one another sequentially - the resulting ssDNA or ssRNA would then have to be converted into a stable ds-form via the usual means, or otherwise protected in one of the usual ways. Codon space could be expanded by pre-loading carrier molecules with more than one type of triplet per carrier (biased towards whatever codon frequencies are prominent in the organism of choice) although this in no way resolves the random nature of the actual codon use within the resulting nucleotide sequence. The issue of amino acid coupling selectivity is pretty hairy - the best I could think of on a short notice is to have the receptor sites for individual aa's arranged in order of dropping selectivity -- however there is still the matter of shape/property similarities throwing wrenches into the works. An alternative would be a series of binary gates working on an exclusion principle. As to practicality of this kind of stuff - I am not sure; I can imagine an application similar to nano-scale multiparallel pyrosequencing: an unknown protein would be broken down into peptides via nonselective protease of some sort and then relatively short individual peptides are 'sequenced' in parallel, producing short DNA sequences that would later be complemented to dsDNA and allowed to cross-anneal and self-assemble via overlaps, similar to gapped gene assembly from short synthetic fragments (that first protease better be *really* non-specific!). At the end one could sequence the resulting long DNA to see what the original protein was like. A. On Tue, Sep 7, 2010 at 8:35 AM, David Schuller dj...@cornell.edu wrote: On 09/06/10 21:36, Jacob Keller wrote: Dear Crystallographers, does anyone know of any conceptual reason why a reverse translatase enzyme (protein--nucleic acid) could not exist? I can think of so many things for which such an enzyme would be helpful, both to cells and to scientists...! Unless there is something I am missing, it would seem to me conceptually almost impossible that it *not* exist. See: The RNA/Protein Symmetry Hypothesis: Experimental Support for Reverse Translation of Primitive Proteins Masayuki Nahimoto, J. Theor. Biol. (2001) 209, pp 181-187. In which Nahimoto proposes such a system, and additionally proposes that it actually existed early in the development of life on this planet. Reasons why it could not exist - No. Reasons why it would be very difficult - yes. And plenty of reasons why Nahimoto is probably wrong about it having actually existed: There is absolutely no evidence presented that such a system was ever in operation in the history of life on this planet. Current theories such as the RNA World are much more likely explanations for how life as we currently know it may have developed from a pre- biotic state. DNA replication, DNA=RNA transcription, and RNA=Protein translation all depend on nucleic acid base pairing for part of their specificity. It truly is the secret of life. And it would not be especially helpful in Protein=RNA reverse translation. Forward translation takes place in the ribosome, but extra specificity is smuggled in via a large set of tRNAs and tRNA charging enzymes, in reactions which took place beforehand, which are then made use of through the base-pairing codon:anti-codon recognition. Reverse translation would most definitely not be running forward translation in reverse; the specificity cannot be handled ahead of time, it needs to be available at the site of reverse translation itself when each successive peptide residue is presented. Progressivity: If different recognition sites are swapped in, this has to be done while keeping place in both the protein chain and in the growing nucleotide chain. Possibly the protein chain might be cleaved during the process. The chemistry and geometry of peptide residues is far more variable
[ccp4bb] How to switch objects in Pymol animation
Sorry for the off-topic question. I am trying to make a complex animation using new movie-making features of MacPymol 1.2. The video tutorials that are available to power users explain how to use the timeline, however, they do not tell how to switch objects during the animation. Is it possible when using the timeline? I know plugins emovie and slerpy allow to switch/fadeout objects, but they do not move them relative to each other that is so easy to do in the new Pymol. Alex
Re: [ccp4bb] Jobs
Are the salaries compared in orders of magnitude? Or you mean other pays? On Sep 30, 2009, at 8:30 PM, William Scott wrote: I always Look on the Bright Side of Life, so I take a certain solace in the fact that while this may be true, most postdoc positions pay about as well as my job, if not better. On Wed, September 30, 2009 5:52 pm, Artem Evdokimov wrote: I personally am not peeved at all, but I am getting concerned that most of the jobs advertised are postdocs. Mighty and mysterious market forces have cut very deep into the heart of structural science and this is not good. Artem