grandrea
commented
1 month ago If you are interedted in PPIs, you should utilize xiFDR, filter on both residue pair level (5% for example) and protein pair level (also 5%) and then report all links passing fdr to have appropriate statistical control. See for lenz et al nat comms 2021 for more on this.
You then can deposit mzIdentML in PRIDE/ProteomeXChange and attatch the links csv as supplementary data to your paper.
Filtering based on score post-fdr sort of defeats the purpose of a target-decoy fdr in the first place...
On Fri, 11 Oct 2024, 18:37 Luca, @.***> wrote:
Is it necessary to filter out the results based on the peptide matching score after performing the estimation of FDRs?
For example, would you avoid reporting a PPI based on a small matching score value for the detected crosslinked peptide? If so, what would be your threshold for deciding which crosslinks are worth reporting and which are not?
— Reply to this email directly, view it on GitHub https://github.com/Rappsilber-Laboratory/xiFDR/issues/49, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKR4YA72ELE3WUP47NSEKATZ275DFAVCNFSM6AAAAABPZIL32GVHI2DSMVQWIX3LMV43ASLTON2WKOZSGU4DCNZXGYZTANQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Is it necessary to filter out the results based on the peptide matching score after performing the estimation of FDRs?
For example, would you avoid reporting a PPI based on a small matching score value for the detected crosslinked peptide? If so, what would be your threshold for deciding which crosslinks are worth reporting and which are not?