PeptideProphet: You need to provide a protein database

[jamesknight21]

Dear members of the Philosopher team,

Firstly, thank you making such a great software. I am very excited to use philosopher on our linux cluster.

I am in the testing phase of philosopher, and I keep encountering a fatal error (You need to provide a protein database) in peptide prophet.

As a sanity check, my workplace initialisation went fine, so did the search using MSFragger. My configuration file (philosopher.yml) is also probably correct, as I can use the same configuration (yml + fas file) to properly execute searches of up to two workplaces (i.e. folders containing the mzML file). However, as soon I as try three or more, I encounter that fatal error.

Any advise? Please let me know any additional detail that will help in better communication of the problem.

Sincerely,
James

[sarah-haynes]

Hi James, are you running each step (workspace creation, database search, etc.) using the pipeline command? You should be able to search any number of spectral files in a given analysis, but each analysis should be performed with a separate philosopher.yml file. The pipeline tutorial has an example of how multiple files can be organized and searched in a single analysis.

Sarah

[prvst]

Hi @jamesknight21. Thank you for the kind words. Could you share your YAML file with us? Also, for future reference, if you find an issue or a problem with the program, please use the issue tracker, instead of the discussion board.

[jamesknight21]

Dear Sarah and Felipe,

@hayse1 I am using the pipeline command. I am using philosopher workspace --init and philosopher data --custom my_uniprot.fasta --contam to setup my workplace and protein database folder. Additionally, these are all label free runs. The following is my folder structure:

test/
|-- test0/
..|-- test0.mzML
|--  test1/
..|--test1.mzML
|-- test2/
..|--test2.mzML
|-- test3/
..|--test3.mzML
|-- 2021-06-15-decoys-contam-my_uniprot.fasta.fas
|-- philosopher.yml

My execution order is:
/test/ $ philosopher workspace --init
/test/ $ philosopher database --custom ~/fasta/my_uniprot.fasta --contam
/test/ $ philosopher pipe-line --config ./philosopher.yml test0/ test1/ test2/ # This raises the fatal error.
/test/ $ philosopher pipe-line --config ./philosopher.yml test0/ test1/ # Any combination of the two will successfully complete.

Just as a recap, peptideprophet is raising the fatal error (You need to provide a protein database). The MSFragger based closed search using the .fas file has ALWAYS completed successfully.

EDIT: @prvst @hayse1 Issue is resolved!! It had to do with the concurrent option for peptideprophet in the yml file. Nonetheless, thank you to your team for the wonderful software (also the innovative algos in MSFragger), and for thoughtful support!

Lastly, @prvst @hayse1, which parameters are you using in MSCONVERT for the conversion of thermo raw files into mzML?
My current parameters are: --mzML --zlib --filter "peakPicking vendor". What parameters do you recommend for PASEFDDA data from Bruker TIMSTOF instrument? FYI, I am doing this on command line using their docker image.

Sincerely,
James

[sarah-haynes]

Hi James, we don't recommend converting DDA PASEF .d files from Bruker, the raw .d file should be used directly. We have some notes about spectral file formats for FragPipe, and you should still use .d if you are searching with MSFragger with Philosopher pipeline. If you do need to quantify PASEF data, use FragPipe.

For converting Thermo .raw files, --mzML --64 --zlib --filter "pickPicking true 1-" *.raw is equivalent to the GUI options shown in our conversion tutorial.

Sarah

[jamesknight21]

Hi Sarah,

Thank you for the reply.

I have gone through the conversion tutorial. However, I have a lingering question. I wish to only use philosopher with MSFragger and Ionquant. This is primarily because of restrictions associated with our Linux cluster.

Hence, what is your recommendation for analyzing DDA PASEF .d files using Philosopher/MSFragger on a Linux (ubuntu) computer? Apologies if my question is naïve, currently I am finishing the implementation of philosopher with thermo's .raw on our Linux cluster. Next week, I plan to implement Philosopher/MSFragger based analysis of DDA PASEF .d files.

Sincerely,
James

[prvst]

Thanks again, and glad to hear you found it. The concurrent option sometimes hangs up like that, I will hopefully have a fix for that in the next release.

[jamesknight21]

Good to know... thank you Felipe!