Troubleshooting Low Protein Identifications in DIA-MS Data Despite Good Sample Quality

[lizexhusselmann]

I would like to get advise on the following analysis:
Problem 1:
Instrument: Bruker timsTOF Pro (DIA mode).
Sample preparation: Protein extractions from SDS-PAGE gels appeared intact and of good quality for both Control and Fungal-treated groups.
Initial QC: Service provider confirmed that TIC profiles were comparable and visually acceptable for both datasets.
Data processing: DIA-NN was used to generate report_pg_matrix.tsv.
Problem 1: Only ~315 proteins were identified in the treated samples, versus ~4,800 in controls.
Problem 2: When importing the matrix into FragPipe Analyst, an excessive number of proteins show imputed intensity values, suggesting that many identifications are missing or below detection limits.

[vdemichev]

Hi,

I will need full DIA-NN logs and DIA-NN's PDF report ('runs' one) to comment on the issue. Please also see suggestions here https://github.com/vdemichev/DiaNN#troubleshooting.

Best,
Vadim

[lizexhusselmann]

Dear Vadim

Thank you very much for the response. See the attached docs as requested.
Kind regards
Lizex

report.log.txt
report_runs.pdf
report_trends.pdf

[vdemichev]

I don't see any issues with mass calibration, RT distribution also similar. I can only suggest to run this also with InfinDIA in DIA-NN 2.6, try different stuff that may reveal a problem with the sample: modifications, up to 2 missed cleavages, semi-tryptic search (each separately). Maybe look at which peptides do get identified, what's the common pattern. Also address the incorrect mass accuracy setting warning.

[lizexhusselmann]

Hi Vadim

Thank you for the suggestions. I'll address these and keep you posted.

Kind regards

Lizex