Not sure if I'm running this correctly. All QValue and PepQValue are 0

[ttessie2]

I've just starting running MSGF+. I'm just using some publicly available yeast data from MASSIVE. I used philosopher to create my target-decoy database. Things seem to be running without any errors however when I look at the tsv file all the QValues are 0. I double checked the -decoy parameter to make sure it matched what my decoys are appended with and that didn't solve anything. If I search through the tsv file I find rev decoy hits in there. Where could this source of error be coming from?

[FarmGeek4Life]

Because you used philosopher to create your target-decoy database, you are probably searching with -tda 0. If you use -tda 0, the QValues (and PepQValues) are not calculated; MS-GF+ does not automatically check for decoy prefixes in the .fasta file.

What should work for this is the following:

• Make a copy of the .fasta file, but have it end with .revCat.fasta instead of .fasta
• Run the search with -tda 1.

What this does is bypass the MS-GF+ decoy creation process, while still performing a full target-decoy search.

[ttessie2]

Thank you for the reply! I originally had it set to -tda 0 but the error you get says:

Error while indexing: 2020-04-08-decoys-contam-UP000002311.revCat.revCat.fasta (too many redundant proteins) If the database contains forward and reverse proteins, run MS-GF+ (or BuildSA) again with "-tda 0" If the decoy protein names do not start with XXX either rename them, or use the -decoy switch

After reading that I switched it to '-tda 1' and added '-decoy rev' because the database contains forward and reverse proteins and they do not start with XXX. Looking at the decoys they are appended rev_ but in my command line I only put rev. Would this make a difference??

For clarity this is what I entered originally. This ran but the QValues were 0. C:\MSGF+>java -Xmx3500M -jar MSGFPlus.jar -s C:\TPP\data\params\WT_Rep_1_Resp_Prot.mzML -d C:\FragPipe_Skyline\Philosopher\2020-04-08-decoys-contam-UP000002311.fa -inst 1 -t 20ppm -ti 1,2 -ntt 2 -tda 1 -decoy rev -o demo.mzid

[FarmGeek4Life]

Do not enter the ".revCat.fasta" on the command line for MS-GF+; give it the ".fasta" file, and have the ".revCat.fasta" file in the same directory. MS-GF+ will automatically find the ".revCat.fasta" file and use it instead of generating it.

[FarmGeek4Life]

And, let me look at the code a little; it's possible that there is some automatic handling of some of this that I don't remember.

[ttessie2]

Okay, so should I have -d database.fasta -tda 1 -decoy rev? So when the program runs it will look for the .revCat.fasta file within the directory? And I shouldn't have the target database and reverse database within the same file?

[FarmGeek4Life]

Okay, so should I have -d database.fasta -tda 1 -decoy rev? yes So when the program runs it will look for the .revCat.fasta file within the directory? yes And I shouldn't have the target database and reverse database within the same file? For the database.fasta file - it may not matter (but it definitely needs the target database). The database.revCat.fasta file needs to have both the target and reverse/decoy database, as one file. (the name "revCat" is just a shortened form of "reverse concatenated", meaning it has both target and decoy hits.)

[ttessie2]

Okay, I'm running this now so I'll see how it goes. When you say the .revCat.fasta should be in the same directory. Are you referring to the same directory the fasta.db file?

[FarmGeek4Life]

database.fasta and database.revCat.fasta need to be in the same directory, e.g. on Windows that might be: C:\msgf\database.fasta C:\msgf\database.revCat.fasta

[ttessie2]

Thanks for the help! I have it working now. Much appreciated! I am new to MS analysis so this has been great. Last question if you don't mind my asking. What is your preferred next step for protein level analysis? I am more familiar with the TPP pipeline using iprophet -> proteinProphet. It looks like MSGF+ just gives PSM and peptide level analysis, is that correct?

[alchemistmatt]

Correct: MS-GF+only identifies peptides and reports the proteins that they're associated with. Some options for protein rollup are IDPicker and InfernoRDN. There are also several commercial tools that do a great job, including Scaffold

I suggest IDPicker, since it supports protein parsimony and combining multiple datasets. In contrast, InfernoRDN, just supports protein rollup on a single dataset at a time. IDPicker should be able to read the .mzid files created by MS-GF+.

If you want to perform quantitation (using Selected Ion Chromatograms of the MS1 parent ions), you'd have to analyze your data with MASIC then merge the MASIC results with MS-GF+ using MASIC Results Merger.

[ttessie2]

Great, thank you!

[Jokendo-collab]

I think MS-GF+ developers should make the target-decoy search a default with tda 1 not tda 0

On Sat, Apr 11, 2020 at 2:39 PM ttessie2 notifications@github.com wrote:

Closed #98 https://github.com/MSGFPlus/msgfplus/issues/98.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/MSGFPlus/msgfplus/issues/98#event-3224019687, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGJ34O5GFSG6YSYDYW2LXRDRMBQIBANCNFSM4MFSZDWQ .