Setting Precursor and Protein PEP in DIA-NN

[barny116]

Hi Vadim,
Hope this message finds you well.
I am currently searching in DIA-NN using two proteomes- mouse and human, making the search space rather large. Do you have any advice on how to make the search more stringent? Is it possible to set a protein q-value cutoff by both experiment and run in DIA-NN? Currently, I use a 1% precursor and protein group FDR. Also, can I set a precursor and protein PEP (Posterior Error Probability) in DIA-NN?
Thanks!

[Vadim Demichev]

Hi Barny,

I assume you use MBR and search with the default FDR filtering set in the GUI (1%), in that case all you need to do is to filter the final report (in .parquet format) using Lib.PG.Q.Value <= 0.01 and PG.Q.Value <= 0.01.

Yes, you can use PEP and PG.PEP columns of the main report for extra stringency.

Best,
Vadim

[barny116]

Hi Vadim, Thanks so much for your help. In my understanding, Lib.PG.Q.Value <= 0.01 and PG.Q.Value <= 0.01 would only filter on the protein level. Is this correct? Do you also recommend also filtering on the precursor level, adding in Lib.Q.Value <= 0.01 and Q.Value <= 0.01? Previously, I was filtering using the global.Q.Value <= 0.01 and Q.Value <= 0.01. I now realize I should have been taking to account the fact that I am using MBR. Also, I am unable to find PG.PEP in my DIA-NN result file. Why would this be? Thanks!

…

On Mon, Jul 1, 2024 at 11:40 AM Vadim Demichev ***@***.***> wrote: Hi Barny, I assume you use MBR and search with the default FDR filtering set in the GUI (1%), in that case all you need to do is to filter the final report (in .parquet format) using Lib.PG.Q.Value <= 0.01 and PG.Q.Value <= 0.01. Yes, you can use PEP and PG.PEP columns of the main report for extra stringency. Best, Vadim — Reply to this email directly, view it on GitHub <#1070 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A62XJWBIEAHK2M56MHNPF4LZKGBBNAVCNFSM6AAAAABKF43TY6VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMBQGYYDENZXGM> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[Vadim Demichev]

Hi Barny,

If you use MBR and analyse with the default FDR setting of 1% in the GUI, then the Q.Value <= 0.01 and Lib.Q.Value <= 0.01 filters are applied automatically. Same if you also 'manually' reproduce MBR by first creating an empirical DIA-based library and subsequently analysing with it.

PG.PEP is saved in the .parquet report, please use it rather than the .tsv.

Best,
Vadim

[barny116]

That makes more sense! I have been running DIA-NN 1.8.1 for some time now, so I was filtering previously on the report.tsv.
I checked again the report.tsv and still didn't see PG.PEP. I will switch over to 1.9!