Hi David, Thanks for pointing out the wrong mass - there were a couple of others as well. So I've got quite a lot further with the analysis, but I'm not convinced that it's working as it should. Maybe it's easiest if I detail the steps I've gone through and ask some questions as I go. It would be great if you could answer these / annotate with any other thoughts you have:
Step 1: Search using standard 14N labelled masses. The 15N search seems to be fairly useless (see later). No variable modifications in search, because otherwise ASAPRatio complains and fails to run in 'static' mode. Step 2: InteractParser $intout $in Step 3: PeptideProphetParser $intout DECOY=XXX_ NONPARAM Note I get a lot of model failures for this data. I get less model failures when I use fully parametric modelling, but then the decoy search hits are not properly taken into account meaning that in the final prot.xml file I have single -protein protein groups consisting only of a decoy hit. So I continued with the semi-parametric options. Step 3: ASAPRatioPeptideParser $in -lACDEFGHIKLMNPQRSTVWY -r0.5 -S -mA72.04581R160.1359N116.0603D116.0356C161.0393E130.0513Q130.076G58.03016H140.085I114.0928L114.0928K130.1124M132.0492F148.0771P98.06146S88.04073T102.0564W188.0967Y164.072V100.0771 So the -m string specifies the modified masses (ie 15N masses) for comparison with the 14N search results. This doesn't work the other way round (taking the 15N search results and using the 14N masses in the -m string. This is because of cysteine carbamylation, meaning that ASAPRatio sees the C mass as being a 'heavy' static modification, and the other residues as being 'light' modifications. This results in an error message that there are a mixture of heavy and light modifications). Step 4: ProteinProphet $light $out ASAP_PROPHET I get ratios for reasonable numbers of proteins - but unfortunately I'm not convinced these are reliable. To take one example: > L/H ratio given as 2.98 > Based on peptide: CTTSAAATSTSSGR > Looking at the details in the viewer I see "light +2 m/z 679.3" and "heavy +2 m/z 679.8" > There are 17 N atoms in the peptide, 16 with one neutral loss. m/z for the mass difference between light and heavy in the latter case will then be 8. So ASAPRatio should be looking for a mass of about 687. > When I look in the 3D data viewer there are candidate peaks in this region. I also tried xpress and got quantification that were quite different (L/H is 0.66 in this example). The trouble with the xpress output is that there doesn't seem to be a way to click through to the underlying data for inspection. In fact using Petunia I see no way of finding out what peptide or heavy peak identification the ratio is based on. Presumably this information is buried somewhere in the prot.xml / pep.xml files? So my questions are: 1. Should -r be bigger to allow ASAPRatio to find the correct heavy peak? 2. Why does ASAPRatio accept a heavy peak that is so obviously the wrong mass given the static modifications? It seems to be completely ignoring them! 3. Ideally I would use the 15N search results to validate the identity of the heavy peak where possible. Is there a way to do this other than looking it up manually? I've uploaded a pdf with the above example, as well as some data files to: https://we.tl/t-uaqpZ4gy7J Many thanks for the help! Alastair On Wednesday, 26 August 2020 at 22:47:57 UTC+2 David Shteynberg wrote: > Hello Alastair, > > Unless there is a mistake, I think the N15 mass string should be: > > > > -mA72.03415R160.10111N116.04293D116.02694C161.0307E130.04259Q130.05858G58.02146H140.05891I114.08406L114.08406K130.09496M132.04049F148.06841P > 98.05276S88.03203T102.04768W188.07931Y164.06333V100.06841 > > > See the following post: > > https://groups.google.com/g/spctools-discuss/c/O-Ude_j6Nss/m/6Tl2Q-_hAAAJ > > However, I am not detecting correct results in your heavy 'H' search > results. I would start your troubleshooting there, beginning with the > heavy mass of proline. > > Sorry this has taken me some time to get to and keep me posted to your > progress. > > Cheers, > -David > > > > > > > On Wed, Aug 26, 2020 at 2:49 AM 'Alastair Skeffington' via > spctools-discuss <spctools...@googlegroups.com> wrote: > >> Hi David, >> >> Did you manage to grab the data before the link expired? >> >> Thanks, >> Alastair >> >> On Monday, 17 August 2020 at 21:23:55 UTC+2 Alastair Skeffington wrote: >> >>> Hi David, >>> >>> Ah sorry - my mistake. Here's a new link with a tarball including the >>> mxXML files. >>> >>> https://we.tl/t-UnX74n4qod >>> >>> Many thanks! >>> Alastair >>> >>> >>> On Monday, 17 August 2020 20:51:54 UTC+2, David Shteynberg wrote: >>> >>>> Hello Alastair, >>>> >>>> I downloaded the file you shared with me, however, there were no mzML >>>> files to match the pepXML files so I couldn't actually try the analysis. >>>> I will make another attempt if you can provide the mass spec data. >>>> >>>> Thanks, >>>> -David >>>> >>>> On Mon, Aug 17, 2020 at 11:21 AM 'Alastair Skeffington' via >>>> spctools-discuss <spctools...@googlegroups.com> wrote: >>>> >>> Hi David, >>>>> >>>>> The link to the data has now expired, Let me know if you are still >>>>> willing to have a look and I can send it again. >>>>> >>>>> Otherwise maybe you could briefly describe the process you would go >>>>> through? >>>>> >>>>> Many thanks, >>>>> Alastair >>>>> >>>>> On Thursday, 6 August 2020 22:17:00 UTC+2, Alastair Skeffington wrote: >>>>>> >>>>>> Hi David, >>>>>> >>>>>> Many thanks for your reply. >>>>>> >>>>>> So it any peptides with modifications will simply be ignored for >>>>>> quantification and I can ignore the warning message? >>>>>> >>>>>> Yes - each search was either light or heavy as defined in the static >>>>>> modifications. >>>>>> >>>>>> And the -r parameter is then the window to search in the RT >>>>>> dimension? Because the command line option says 'range around precursor >>>>>> m/z >>>>>> to search for peak' I assumed this was for the m/z dimension. I thought >>>>>> that the shift of the peak would also mostly be in the m/z dimension - >>>>>> but >>>>>> I'm no mass spectrometrist! >>>>>> >>>>>> I would be amazing if you had a moment to have a look at the data - >>>>>> thanks very much for offering. I've put two example pairs of files here: >>>>>> https://we.tl/t-cVDD1MVDOr >>>>>> >>>>>> For each sample there is a light 'L' version of the search results >>>>>> and a heavy 'H' version. I've also included the search database (based >>>>>> on >>>>>> some PacBio data some I'm afraid it's quite big with a lot of isoforms). >>>>>> >>>>>> Many thanks, >>>>>> Alastair >>>>>> >>>>>> On Wednesday, 5 August 2020 20:37:16 UTC+2, David Shteynberg wrote: >>>>>>> >>>>>>> Hi Alastair, >>>>>>> >>>>>>> If your search results are either all heavy or all light (not >>>>>>> variable mod searched) then you should also use option -S. >>>>>>> >>>>>>> 1). You cannot specify anything but single amino acids in this >>>>>>> string. Your quantitation will be based on peptides without PTMs in >>>>>>> this >>>>>>> dataset. >>>>>>> >>>>>>> 2). -r8 is a MUCH too wide window to recover the MS1 signal in >>>>>>> RTspace. The lower this number the more selective the tool is at >>>>>>> isolating >>>>>>> your target signal. With -r8 you will not be quantifying the correct >>>>>>> signal, unless you have a very bare sample. >>>>>>> >>>>>>> If you are able to share this data I can try running it to help you >>>>>>> optimize your settings. >>>>>>> >>>>>>> Thanks, >>>>>>> -David >>>>>>> >>>>>>> On Wed, Aug 5, 2020 at 11:02 AM 'Alastair Skeffington' via >>>>>>> spctools-discuss <spctools...@googlegroups.com> wrote: >>>>>>> >>>>>>>> Hello, >>>>>>>> >>>>>>>> I'm trying to run a 14N/15N labelling experiment through ASAPRatio. >>>>>>>> I've been running the first steps like this - here for the results of >>>>>>>> a >>>>>>>> database search with heavy masses: >>>>>>>> >>>>>>>> InteractParser sample_interact.pep.xml sample.pep.xml >>>>>>>> >>>>>>>> PeptideProphetParser sample_interact.pep.xml >>>>>>>> >>>>>>>> RefreshParser sample_interact.pep.xml >>>>>>>> ./EhuxAllproteins_MCC_decoy.fasta >>>>>>>> >>>>>>>> ASAPRatioPeptideParser sample_interact.pep.xml >>>>>>>> -lACDEFGHIKLMNPQRSTVWY -r8 >>>>>>>> -mA72.0779R160.1857N116.1026D116.0874C104.1429E130.1140Q130.1292G58.0513H140.1393I114.1576L114.1576K130.1723M132.1961F148.1739P99.1152S89.0773T102.1038W188.0793Y164.0633V101.1311 >>>>>>>> >>>>>>>> At this point I get a warning: >>>>>>>> >>>>>>>> WARNING: Found more than one variable mod on 'M'. Please make sure >>>>>>>> to specify a heavy mass for this residue >>>>>>>> >>>>>>>> So I have two questions: >>>>>>>> >>>>>>>> 1) How do I specify the heavy mass for oxidised methionine? Mox ? >>>>>>>> And is phosphorylated serine coded Sp ? >>>>>>>> >>>>>>>> 2) I've used -r8 instead of the default 0.5. My reasoning is that a >>>>>>>> medium sized heavy peptide could easily differ from the 14N >>>>>>>> counterpart by >>>>>>>> 16 Da. Assuming charge +2, then using a m/z range of 8. Does this >>>>>>>> sound >>>>>>>> remotely sensible? >>>>>>>> >>>>>>>> Many thanks! >>>>>>>> Alastair >>>>>>>> >>>>>>>> -- >>>>>>>> You received this message because you are subscribed to the Google >>>>>>>> Groups "spctools-discuss" group. >>>>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>>>> send an email to spctools...@googlegroups.com. >>>>>>>> To view this discussion on the web visit >>>>>>>> https://groups.google.com/d/msgid/spctools-discuss/4853fda5-cc2f-48e2-a9e2-2ae93225b288o%40googlegroups.com >>>>>>>> >>>>>>>> <https://groups.google.com/d/msgid/spctools-discuss/4853fda5-cc2f-48e2-a9e2-2ae93225b288o%40googlegroups.com?utm_medium=email&utm_source=footer> >>>>>>>> . >>>>>>>> >>>>>>> -- >>>>> You received this message because you are subscribed to the Google >>>>> Groups "spctools-discuss" group. >>>>> To unsubscribe from this group and stop receiving emails from it, send >>>>> an email to spctools...@googlegroups.com. >>>>> >>>> To view this discussion on the web visit >>>>> https://groups.google.com/d/msgid/spctools-discuss/3f3e8d01-8fd7-4d40-acde-1b7147769108o%40googlegroups.com >>>>> >>>>> <https://groups.google.com/d/msgid/spctools-discuss/3f3e8d01-8fd7-4d40-acde-1b7147769108o%40googlegroups.com?utm_medium=email&utm_source=footer> >>>>> . >>>>> >>>> -- >> You received this message because you are subscribed to the Google Groups >> "spctools-discuss" group. >> > To unsubscribe from this group and stop receiving emails from it, send an >> email to spctools-discu...@googlegroups.com. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/spctools-discuss/c46993a6-b014-4790-b6bd-5afa5810e550n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/spctools-discuss/c46993a6-b014-4790-b6bd-5afa5810e550n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "spctools-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discuss+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/e4cce92f-9086-4d57-8853-23d73abbab8fn%40googlegroups.com.
