Spectral library generation

[Srinivas Srikanth]

Hi Phil,

I have a 15 hek files searched with Mascot, X!Tandem and MSGF+ separately. I am trying to create a library using TPP pipeline and SpectraST.

I have convert the results individually to pepxml using Mascot2XML, Tandem2XML and Idconvert and then apply peptideprophet on each file. My question is about the next step,

Should I be merging individual search engine results using iprophet and to make three combined files and then merge them again using iprophet?

Or I merge each run from three searches together using iprophet and then merge the resulting 15 files together to apply ProteinProphet or Mayu ?

Srikanth

[Eric Deutsch]

Hi Srikanth, I suggest after your step “I have convert the results individually to pepxml using Mascot2XML, Tandem2XML and Idconvert and then apply peptideprophet on each file” (which is good), the next step is merge the 3 PeptideProphet outputs into a single PepXML file with iProphet. Then build your spectral library from that one file.

 

Regards,

Eric

--
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/08ebe555-a53b-4c20-8ac8-14709b4f2d49o%40googlegroups.com.

[Srinivas Srikanth]

Hi Eric,

Thanks for the suggestion. I used iprophet and merged the files successfully, followed by Mayu to calculate the proteinFDR. I also was able to create a spectral library using spectraST.

The place where I get an error is when I create consensus library and try to define the user modifications. I only want Carbamidomethylation at C and Oxidation at M.

I created a text file with the following lines:

# spectrast.usermods

M|+16|

C|+57|

And I run the following command :

 spectrast -cAC -Mspectrast.usermods -cN/speclib_modfiltered -cICID_QTOF -cDUniprot_RRP_for_Spectrast.fasta -cd speclib.splib

It created a library successfully but throws the following error:

GENERAL: Illegal user-defined modification token: SpectraST . It must be of the form A[...], where ... does not start with a number. Ignored.

And when I convert the results to text, I see other modifications as well (deamidation, acetylation, etc). Can you please tell me where I am going wrong.

Srikanth

On Wednesday, June 10, 2020 at 4:28:54 AM UTC+10, Eric Deutsch wrote:

Hi Srikanth, I suggest after your step “I have convert the results individually to pepxml using Mascot2XML, Tandem2XML and Idconvert and then apply peptideprophet on each file” (which is good), the next step is merge the 3 PeptideProphet outputs into a single PepXML file with iProphet. Then build your spectral library from that one file.

 

Regards,

Eric

 

 

From: spctools...@googlegroups.com <spctools...@googlegroups.com> On Behalf Of Srinivas Srikanth
Sent: Monday, June 8, 2020 9:43 PM
To: spctools-discuss <spctools...@googlegroups.com>
Subject: [spctools-discuss] Spectral library generation

 

Hi Phil,

 

I have a 15 hek files searched with Mascot, X!Tandem and MSGF+ separately. I am trying to create a library using TPP pipeline and SpectraST.

I have convert the results individually to pepxml using Mascot2XML, Tandem2XML and Idconvert and then apply peptideprophet on each file. My question is about the next step,

Should I be merging individual search engine results using iprophet and to make three combined files and then merge them again using iprophet?

Or I merge each run from three searches together using iprophet and then merge the resulting 15 files together to apply ProteinProphet or Mayu ?

 

 

Srikanth

--
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.

[Aaron Robinson]

Hi Srikanth,

You shouldn’t need a spectrast usermods file for those common mods.  Spectrast is just reading your search results so if there are other modifications present which it is reading, they are there because you’ve searched for them using tandem/mascot/etc.

I suggest re-searching your data only for the mods of interest and going through the pipeline again to ensure peptideprophet modeling is correct.   

Hope this helps,

-Aaron

Sent from my iPhone

On Jun 10, 2020, at 12:37 AM, Srinivas Srikanth <srikan...@gmail.com> wrote:



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/217c62c8-18ee-411f-a271-20300fe8ecffo%40googlegroups.com.