first search much higher with 1.9.2 and MBR using speclib prediction with supplied library

vdemichev / DiaNN

DIA-NN - a universal automated software suite for DIA proteomics data analysis.

Other

284 stars 53 forks source link

first search much higher with 1.9.2 and MBR using speclib prediction with supplied library #1281

Open jessegmeyerlab opened 3 days ago

jessegmeyerlab commented 3 days ago

report.log.txt

library used is here: https://zenodo.org/records/12802609

data from massive associated with this preprint - https://www.biorxiv.org/content/10.1101/2024.07.23.604834v1.full

vdemichev commented 3 days ago

Hi Jesse,

Thank you for spotting this. I have now added a note in the docs and 1.9.2 release notes that predicted library generation should always happen separately from raw data analysis. There will be a warning in the next release. Specifically, the problem with on-the-fly prediction occurs in 1.9.2 if the retention times in the library are significantly different from the iRT scale.

Best, Vadim