Re: [Nessy-users] problems

Wed, 16 Feb 2011 13:28:11 -0800 

Hi Mike
I am usually fitting to the all residues at once. The fast exchanging models are converging much faster than the slow exchanging (models 3 and 5). 


Do you have data recorded at two fields? I think using the slow 
exchanging models should only be done if you have data at two fields, as 
you might pick up false results due to experimental error and get 
completely wrong populations.



If you say that the program hangs, does this mean noting is calculated 
anymore (no output in the panel) ? Or is the program not reacting anymore?



If you are using all 5 models, calculation for 100 residues (with 
exchange expected for 30 residues) might take you several days.


Cheers
Michael


Am 17.02.2011 07:40, schrieb Michael Osborne:

Hello Michael,

This seems to have done the trick. I tried 10 residues withfit to 
models 1-3 and model selection using AICc.
However, I have had some cases where the programme has hung: this has 
been when it cannot converge (I guessing) when trying to fit  the data 
(normally model 3 or higher), no errors are given the programme just 
hangs..
It seems that the offending residue can be fit if I do just that 
residue, so maybe it is a fitting problem or maybe just a lot of 
memory. When you fit data are you able to do all residues at one time, 
or do you choose groups??



Thanks for all your help. I look forward to seeing a copy of the 
manuscript when it is accepted!!


Cheers
Mike

On 2/15/11 11:43 PM, Michael Bieri wrote:

I sourced the problem in general, but I am not sure if it fixes it 
completely.



Can you please repeat the calculation using the updated version of 
NESSY (Revision 408).


Cheers


Am 16.02.2011 14:59, schrieb Michael Osborne:

Hi Michael

Just a quick note that in the version 407, the evaluate data, does 
not work (that was why it was so quick I guess). However, it does 
plot all the residues selected: however, I still get the error when 
I use more than one residue:



Unhandled exception in thread started by <class 
func.calc.Run_dispersion_calc at 0x15987750>

Traceback (most recent call last):
  File "/Users/mike/bin/nessy/func/calc.py", line 67, in __init__
    Run(self.main, exp)
  File "/Users/mike/bin/nessy/func/calc.py", line 144, in __init__

    Model_Selection(self.main, 
str(self.main.proj_folder.GetValue()), self.exp_index)
  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
136, in __init__

    self.create_plots()

  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
416, in create_plots

    y_R2 = array(y_R2)
ValueError: setting an array element with a sequence.

Cheers
MIke

On 2/15/11 10:31 PM, Michael Bieri wrote:

Aha, almost my expectations... It seems that mac counts in another 
way than linux and windows. If you open up the dialog to import 
nmrview tables (I just created), you should see Experiment 1 and 
not Experiment 0.... This value can only be changed if you add more 
experiments (green + button in toolbar). It seems as there is an 
indexing problem. And this is why plotting does not work. It likes 
to read the i-th data set (linux and windows) and reads the i-th + 
1 dataset in Mac, which is empty....



How many plots are in your results and R2eff_plots folder? One svg 
or png file should have been created for each residue. I assume one 
is missing?


Cheers

Am 16.02.2011 14:23, schrieb Michael Osborne:

Hi Michael
I have just updated to revision 407:

I noticed that when you click on evaluate data, the "in progress" 
gui does;t come up. I think his might be because it does the 
evaluation a lot quicker (or maybe not at all)



trying 2 models with a single residue gives the following message, 
but the prog keeps running:



/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/matplotlib/axes.py:2381: 
UserWarning: Attempting to set identical left==right results

in singular transformations; automatically expanding.
left=9, right=9
  + 'left=%s, right=%s') % (left, right))


When I then choose 3 models it works, which is a success compared 
to the other version!!!
However there still seems to be a problem when using more than one 
residue at a time. e.g



If I select residues 3-9 for a 2model analysis: the fits are done 
however in this case after the MC analyses the software hangs the 
error is:

IndexError: list index out of range

Unhandled exception in thread started by <class 
func.calc.Run_dispersion_calc at 0x15987720>

Traceback (most recent call last):
  File "/Users/mike/bin/nessy/func/calc.py", line 67, in __init__
    Run(self.main, exp)
  File "/Users/mike/bin/nessy/func/calc.py", line 144, in __init__

    Model_Selection(self.main, 
str(self.main.proj_folder.GetValue()), self.exp_index)
  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
136, in __init__

    self.create_plots()

  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
416, in create_plots

    y_R2 = array(y_R2)
ValueError: setting an array element with a sequence.



I managed to also fit models 1-4 when only fitting a single 
residue at the time. so one residue at a time seems OK, not sure 
when using more than one residue. By the way thanks for your quick 
replies!!

Cheers
Mike

ps it reads in the nmrview data nicely. There is another way to 
save the data from nmrview, but that does not include the residue 
numbers, and basically is the same file I gave you up to (but not 
including) the "residue" column. By the way, when loading the 
nmrview format, I could not select the expt number (it stayed at 
zero)


On 2/15/11 9:51 PM, Michael Bieri wrote:

I could not reproduce this problem. I am running NESSY on Ubuntu. 
I remember having a similar problem in the Global fit function on 
Windows. I tryed fixing up something, maybe you can test it on 
Mac? Just update your NESSY copy using svn up.


Cheers


Am 16.02.2011 13:17, schrieb Michael Osborne:

Hi Michael

I'll send this to you first, and since there are a number of 
problems with the calculations, (most reproducible) I'll let you 
decide if I should send these to the mailing list (I would be 
happy to). Also in response to your equations on the nmrview 
format. I created the table using the relaxation panel (Fit 
table) and saved this. For my case, I used peak number as this 
was just quicker for me, of course for proper analysis its 
preferable to use assignment (which is included under the 
residue column). having an option to use either is good. I could 
not get NMRview to spit out the real T values, but so long as 
you know the values, this is not a problem you just need to 
select the correct column for a T value used in the setup data. 
Of course I can always edit this file or simplify/modify it 
using awk if you like, but the it wouldn;t be universal!
Below are some issues I had using nessy, suffice to say I 
haven't got a model 3 plot yet!!
Thanks and as I said, I'll send the problems to the mailing list 
if you so desire.

MIke


I have encountered a number of issues trying to use NESSY, 
especially during the calculations when running NESSY on my mac 
operating system



I have tried to be methodical in troubleshooting. Here is a 
summary of some of the more reproducible problems I have 
encountered, note all these results were obtained after reading 
a saved project.


A) Tried to fit 5 residues with default NESSY settings:


Managed to fit models 1 2 and 3. Did MC analysis for 1-5 
however, after residue 5 the programme crashes with this error 
in the console:
Python(22151,0xa058c720) malloc: *** error for object 
0x1ed998c0: incorrect checksum for freed object - object was 
probably modified after being freed.

*** set a breakpoint in malloc_error_break to debug

Python(22151,0xb01c9000) malloc: *** error for object 
0x1f51da60: incorrect checksum for freed object - object was 
probably modified after being freed.

*** set a breakpoint in malloc_error_break to debug
Segmentation fault

B) Tried fitting a single residue with default NESSY settings:
did fits to models 1 2 and 3 after MC analysis crashed:
the following error was in the console:



Tue Feb 15 20:16:41 RMN01-IRIC.local Python[22195] <Error>: 
doClip: empty path.
Tue Feb 15 20:16:41 RMN01-IRIC.local Python[22195] <Error>: 
CGContextRestoreGState: invalid context

Bus error

C) Tried B) but with AIC as model selection:

This time an error occurred after the MC analysis (did not 
crash). The error was:
/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/matplotlib/axes.py:2381: 
UserWarning: Attempting to set identical left==right results

in singular transformations; automatically expanding.
left=8, right=8
  + 'left=%s, right=%s') % (left, right))

D) as C) but with only models 1 and 2
This worked

E) As C but went back to AICc slection
This worked

F) added 2 extra residues for analsysis everything else same as E
after MC analysis got the following error (did not crash)
Unhandled exception in thread started by
Traceback (most recent call last):

  File "/Users/mike/bin/nessy/elements/redirect.py", line 89, in 
write

    wx.CallAfter(self.error_message.text.AppendText, (string))

  File 
"/usr/local/lib/wxPython-unicode-2.8.11.0/lib/python2.6/site-packages/wx-2.8-mac-unicode/wx/_core.py", 
line 14634, in CallAfter

    app = wx.GetApp()

  File 
"/usr/local/lib/wxPython-unicode-2.8.11.0/lib/python2.6/site-packages/wx-2.8-mac-unicode/wx/_core.py", 
line 7752, in GetApp

    return _core_.GetApp(*args)
  File "/Users/mike/bin/nessy/func/calc.py", line 67, in __init__
    Run(self.main, exp)
  File "/Users/mike/bin/nessy/func/calc.py", line 144, in __init__

    Model_Selection(self.main, 
str(self.main.proj_folder.GetValue()), self.exp_index)
  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
136, in __init__

    self.create_plots()

  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
416, in create_plots

    y_R2 = array(y_R2)
ValueError: setting an array element with a sequence.



G) tried fitting 20 residues to models 1 and 2 only. The rest 
were left to NESSY defaults:

Same error as above:
error
Unhandled exception in thread started by
Traceback (most recent call last):

  File "/Users/mike/bin/nessy/elements/redirect.py", line 89, in 
write

    wx.CallAfter(self.error_message.text.AppendText, (string))

  File 
"/usr/local/lib/wxPython-unicode-2.8.11.0/lib/python2.6/site-packages/wx-2.8-mac-unicode/wx/_core.py", 
line 14634, in CallAfter

    app = wx.GetApp()

  File 
"/usr/local/lib/wxPython-unicode-2.8.11.0/lib/python2.6/site-packages/wx-2.8-mac-unicode/wx/_core.py", 
line 7752, in GetApp

    return _core_.GetApp(*args)
  File "/Users/mike/bin/nessy/func/calc.py", line 67, in __init__
    Run(self.main, exp)
  File "/Users/mike/bin/nessy/func/calc.py", line 144, in __init__

    Model_Selection(self.main, 
str(self.main.proj_folder.GetValue()), self.exp_index)
  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
136, in __init__

    self.create_plots()

  File "/Users/mike/bin/nessy/curvefit/model_selection.py", line 
416, in create_plots

    y_R2 = array(y_R2)
ValueError: setting an array element with a sequence.




F) Some residues with model 3 took ages (did not converge?) and 
hung up


So unfrotunately at the moment there are a few issues.















_______________________________________________
Nessy-users mailing list
[email protected]
https://mail.gna.org/listinfo/nessy-users














    











					Reply via email to