Configuring FDR filtering in Philosopher the same way as in FragPipe.

[Munchic]

Hello,

I am trying to run MSFragger + FDR filtering and report using Philosopher in a Linux environment and achieve similar results as in FragPipe in Windows. Attached are the comparison of IDs in FragPipe (left half) and in Philosopher (right half). First picture highlights results after MSFragger search, while second highlights - after FDR filtering.

It looks like before filtering, the number of peptides and PSMs are very similar (I am assuming the very small differences are due to difference in calculations in Linux vs Windows). After filtering, the difference is around 3.5% which is much larger. Consistently, Philosopher runs have fewer peptides after filtering.

The configuration for filtering used in FragPipe was --sequential --razor --prot 0.01. For Philosopher, I also turned on those flags and left other parameters at defaults:

FDR Filtering: # Filter
  psmFDR: 0.01 # psm FDR level (default 0.01)
  peptideFDR: 0.01 # peptide FDR level (default 0.01)
  ionFDR: 0.01 # peptide ion FDR level (default 0.01)
  proteinFDR: 0.01 # protein FDR level (default 0.01)
  peptideProbability: 0.7 # top peptide probability threshold for the FDR filtering (default 0.7)
  proteinProbability: 0.5 # protein probability threshold for the FDR filtering (not used with the razor algorithm) (default 0.5)
  peptideWeight: 1 # threshold for defining peptide uniqueness (default 1)
  razor: true # use razor peptides for protein FDR scoring
  picked: false # apply the picked FDR algorithm before the protein scoring
  mapMods: false # map modifications acquired by an open search
  models: false # print model distribution
  sequential: true # alternative algorithm that estimates FDR using both filtered PSM and Protein lists

I would love to know how to set up the filtering parameters exactly the same in both cases. Or do you think the issue could be somewhere upstream Thank you so much for your help!

Khoi

[prvst]

Hi @Munchic, please look at the steps that run before the filtering, like the database search, the peptide validation, and the protein inference. Even if you have the same parameters for the filtering, any differences in the previous steps might give you different outcomes.

[Munchic]

Thanks you for your reply! It does seem like all other parameters for MSFragger, PeptideProphet, and ProteinProphet are the same (or kept as defaults). Attached are the two config files.
FragPipe_and_Philosopher_parameters.zip

On that note, is there a tool to convert FragPipe workflow files into Philosopher yaml to avoid manual copying?

[Munchic]

Hi Felipe, I found out that in one case, the maxppmdiff was set as 2000000, while in another was 20. I'll re-run to see if I still observe the considerable discrepancy and let you know. I am still curious on whether there is an easy way to convert FragPipe workflow files to Philosopher YAML. Thank you!

[prvst]

sounds good.

[Munchic]

Hi Felipe, I am still observing the same behavior that after filtering, Philosopher consistently has 2-4% fewer identified peptides. I am trying to understand why before filtering the numbers are very similar but after they differ considerably. It seems like all the parameters are equivalent, do you think the defaults differ in FragPipe vs Philosopher? Thank you!

[anesvi]

Philosopher pipeline runs PeptideProphet, in FragPipe I think for most workflows we switched to Percolator. Even if using the same tool, PeptideProphet, Philosopher pipeline runs all files together through PeptideProphet whereas in FragPipe each file is processed separately (which is a slightly better strategy). This could be contributing to what you observe Alexey From: Khoi Pham ***@***.***> Sent: Thursday, March 3, 2022 3:41 PM To: Nesvilab/philosopher ***@***.***> Cc: Subscribed ***@***.***> Subject: Re: [Nesvilab/philosopher] Configuring FDR filtering in Philosopher the same way as in FragPipe. (Discussion #322) External Email - Use Caution Hi Felipe, I am still observing the same behavior that after filtering, Philosopher consistently has 2-4% fewer identified peptides. I am trying to understand why before filtering the numbers are very similar but after they differ considerably. It seems like all the parameters are equivalent, do you think the defaults differ in FragPipe vs Philosopher? Thank you! — Reply to this email directly, view it on GitHub<#322 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AIIMM646PL6YKZG4QQ7LRJTU6EPXRANCNFSM5PYBR4YQ>. Triage notifications on the go with GitHub Mobile for iOS<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.******@***.***>> ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues

[Munchic]

Hi Felipe, I am using sequential filter in both (FrapPipe filter set with --sequential, and Philosopher YAML filter set with sequential: true).

That is wonderful news, thank you for the work that you are doing! Are you able to share when approximately we can expect the command line FragPipe and would it be runnable from Linux? Would it include the same set of modules like the GUI (e.g., IonQuant)? Thank you so much!

[fcyu]

Hi @Munchic ,

Please find the pre-release of FragPipe supporting headless from https://www.dropbox.com/s/wj926l2rkirth7w/FragPipe-17.2-build26.zip?dl=1. Here (Nesvilab/FragPipe#560 (comment)) has a brief description about how to run FragPipe in headless. Everything in FragPipe GUI will be available in FragPipe headless (they are basically the same program).

Best,

Fengchao

[Munchic]

Hi Fengchao, thank you for the link and instructions. This is very exciting, and I'm looking forward to the release! Will test it out soon, and hopefully it resolves the differences I observed in Philosopher.

[Munchic]

@fcyu Hi Fengchao. I wanted to follow-up: is there an approximate timeline for when the headless FragPipe will be released? Are the pre-release version and to-be-released version fully compatible? I am asking to make sure workflows or commands will work the same. Thank you again for your work, and I am excited for updates!

[fcyu]

I don't have the exact timeline for releasing. I am also not sure if the pre-release version will be the same as the to-be-released since we are actively improving it.

Best,

Fengchao