Hi Alex, It sounds like perhaps the 2+ model failed the QC check and you got 0 probabilities for all 2+ PSMs. I am not sure why this would be, one possibility is variable modifications that generate homologous top matches. You can try toaddress this with LEAVE or EXPECTSCORE options that ignore the scores from homologous matches. Also, you may try a decoy search and DECOY= and/or NONPARAM options which are helpful at separating the negative model from the positive model when the two overlap.
-David On Wed, Jul 4, 2012 at 8:35 AM, Alex <[email protected]> wrote: > *to complete the picture, here the proteinprophet analysis results* > > * > * > > *ProteinProphet® analysis results* > Version: Insilicos_LabKey_C++ (TPP v4.5 RAPTURE rev 2, Build 201202031108 > (MinGW)) > Analysis Date: 2012-07-04T12:21:56 > > > Source Files: c:/Inetpub/wwwroot/ISB/data/120629YEAST2D/interact.pep.xml > > Number of input spectra with minimum probability 0.05: 0 1+, 0 2+, and 1324 3+ > > Reference Database: c:/Inetpub/wwwroot/ISB/data/dbase/Yeast/YEAST.fasta > Residue substitutions: I -> L > Run Options: > Occam's Razor used: Y > Protein Groups: Y > Peptide degeneracies: Y > Number of Sibling Peptides used: Y > Min peptide probability: 0.20 > Min peptide weight: 0.50 > > Original results written to file: > c:/Inetpub/wwwroot/ISB/data/120629YEAST2D/interact.prot.xml > > *Analysis Iterations* *EM step**number of iterations*Initial Peptide > Weights2NSP Distributions2Final Peptide Weights1 *Learned Number of > Sibling Peptide Distributions* Neighboring bin smoothing: Y *bin number**nsp > range**positive freq**negative freq**positive/negative ratio*00.00 <= nsp > < 0.100.2080.5350.3910.10 <= nsp < 0.250.0650.1420.4620.25 <= nsp < 0.50 > 0.0580.0421.3830.50 <= nsp < 1.000.1240.0741.6741.00 <= nsp < 2.000.125 > 0.0631.9752.00 <= nsp < 5.000.1660.0592.8265.00 <= nsp < 15.000.2000.066 > 3.05715.00 <= nsp < inf0.0540.0192.81 (3.05) > > *Total Number of Contributing Spectrum IDs (100% share): *660.1 > > > > On Wednesday, 4 July 2012 12:32:49 UTC-3, Alex wrote: >> >> Another update to this situation! >> >> I think the problem might actually be in PeptideProphet. I run the same >> dataset using X!Tandem on the GPM.org website, and I do identify doubly >> charged precursor peptide ions. The same dataset only shows triple charged >> precursors, as observed in the pep.shtml file. I select the column: z >> (assumed charge), and ONLY triple charged species appear. >> Here is another twist: When I go to Peptide Prophet Analysis results I >> only see models for 3+, and *No models or data found for charge +2.* >> >> However if I hit on show ALL models, I get the following (which I'm not >> an expert to understand, so hopefully someone can fill me on this >> information): >> >> FINAL 2+ MODEL after 24 iterations: >> number of spectra: 53628 >> using 3+ positive distributions to identify candidates ('-2') above >> background ('0') >> prior: 0.000, total: 0.0 >> X! Tandem (k-score) discrim score [fval] negmean: -1.33 >> pos: (gaussian mean 1.55, stdev 2.01) >> neg: (evd mean -1.31, stdev 1.14, mu -1.82, beta 0.89) >> no. tolerable trypsin term. [ntt] >> pos: (ntt=0 0.000, ntt=1 0.000, ntt=2 1.000) >> neg: (ntt=0 0.000, ntt=1 0.000, ntt=2 1.000) >> no. missed enz. cleavages [nmc] >> pos: (nmc=0 0.878, 1<=nmc<=2 0.122, nmc>=3 0.000) >> neg: (nmc=0 0.409, 1<=nmc<=2 0.591, nmc>=3 0.000) >> accurate mass diff [massd] (offset: -0.500000) >> pos: (massd=-5.000000 0.000000000, massd=-4.000000 0.000000000, >> massd=-3.000000 0.000000000, massd=-2.000000 0.000154395, massd=-1.000000 >> 0.029449633, massd=0.000000 0.390222698, massd=1.000000 0.422314370, >> massd=2.000000 0.128208899, massd=3.000000 0.017069506, massd=4.000000 >> 0.012580499, massd=5.000000 0.000000000) >> neg: (massd=-5.000000 0.000000000, massd=-4.000000 0.000000000, >> massd=-3.000000 0.000000000, massd=-2.000000 0.000173670, massd=-1.000000 >> 0.179740500, massd=0.000000 0.175030116, massd=1.000000 0.164730014, >> massd=2.000000 0.158361143, massd=3.000000 0.159298718, massd=4.000000 >> 0.162665838, massd=5.000000 0.000000000) >> >> >> >> FINAL 3+ MODEL after 30 iterations: >> number of spectra: 30124 >> using training data negative distributions >> prior: 0.023, total: 699.6 >> X! Tandem (k-score) discrim score [fval] negmean: -1.23 >> pos: (gaussian mean 5.15, stdev 2.54) >> neg: (evd mean -1.23, stdev 1.26, mu -1.80, beta 0.99) >> no. tolerable trypsin term. [ntt] >> pos: (ntt=0 0.000, ntt=1 0.000, ntt=2 1.000) >> neg: (ntt=0 0.000, ntt=1 0.000, ntt=2 1.000) >> no. missed enz. cleavages [nmc] >> pos: (nmc=0 0.878, 1<=nmc<=2 0.122, nmc>=3 0.000) >> neg: (nmc=0 0.409, 1<=nmc<=2 0.591, nmc>=3 0.000) >> accurate mass diff [massd] (offset: -0.500000) >> pos: (massd=-5.000000 0.000000000, massd=-4.000000 0.000000000, >> massd=-3.000000 0.000000000, massd=-2.000000 0.000154395, massd=-1.000000 >> 0.029449633, massd=0.000000 0.390222698, massd=1.000000 0.422314370, >> massd=2.000000 0.128208899, massd=3.000000 0.017069506, massd=4.000000 >> 0.012580499, massd=5.000000 0.000000000) >> neg: (massd=-5.000000 0.000000000, massd=-4.000000 0.000000000, >> massd=-3.000000 0.000000000, massd=-2.000000 0.000173670, massd=-1.000000 >> 0.179740500, massd=0.000000 0.175030116, massd=1.000000 0.164730014, >> massd=2.000000 0.158361143, massd=3.000000 0.159298718, massd=4.000000 >> 0.162665838, massd=5.000000 0.000000000) >> >> >> >> >> >> Hope this helps to figure out why I dont get any information for doubly >> charge precursors, which are clearly there on my .wiff data! >> >> Thanks again and hope someone will pick on with this!!! >> >> Alex >> >> >> On Tuesday, 3 July 2012 17:36:50 UTC-3, Alex wrote: >>> >>> I just wanted to add that I checked on previously analyzed samples, and >>> that apparently, with the same settings and experimental conditions, I got >>> the expected distribution of 2+ and 3+ peptides before. This sample was >>> acquired on march of this year, probably with an earlier version of TPP. >>> After that acquisition, I checked my results, and I have only identified 2+ >>> OR 3+ peptides in the pep.shtml files (I didn't realize back then on this >>> peculiar observation). Could someone check if this has happened in your set >>> of results? >>> >>> Thanks! >>> >>> Alex >>> >>> On Tuesday, 3 July 2012 17:03:05 UTC-3, Alex wrote: >>>> >>>> Hi TPP community! >>>> >>>> I have an interesting puzzle that I can not figure out!!! I'm running a >>>> 2D-LC-MS/MS experiment of a yeast lysate (peptides labeled by >>>> dimethylation) on a QqLIT (QTrap 3000) instrument acquiring on IDA mode. >>>> I'm analyzing the data on the TPP using X!Tandem search engine with the >>>> params file attached, and interestingly, ALL the peptides identified on the >>>> pep.shtml file are derived from triply charged precursors!!! (ALL of them). >>>> When I check the prot.shtml file all the identified peptides have the >>>> 3_SEQUENCE nomenclature (my understanding is that the 3 identifies the >>>> charge on the precursor, right?). When I look at the raw data (.wiff) file >>>> on Analyst software, I can easily estimate that most (probably a little >>>> more than half) of the MS/MS was actually performed on CLEARLY doubly >>>> charged precursors, according the the ER-MS (enhanced resolution MS >>>> acquired on the LIT)... so it seem very weird that I dont find a single >>>> peptide to spectrum match (PSM) from a doubly charged precursor. Just >>>> wondering if I might have accidentally changed a parameter on the params >>>> file (.xml). >>>> >>>> Has anybody crossed upon this observation before??? Anybody can help >>>> me??? >>>> >>>> Thanks!!! >>>> >>>> Alex >>>> >>> -- > You received this message because you are subscribed to the Google Groups > "spctools-discuss" group. > To view this discussion on the web visit > https://groups.google.com/d/msg/spctools-discuss/-/wiv8gSP7xLgJ. > > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/spctools-discuss?hl=en. > -- You received this message because you are subscribed to the Google Groups "spctools-discuss" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/spctools-discuss?hl=en.
