Open Bisho2122 opened 8 months ago
@theodev Look at this PR from the point of view of science.
I think we should regenerate all mol_db cache
after deploying this code to production. @lmacielvieira @theodev
The number of annotations will change in the case of reprocessing for existing datasets. This is the problem investigated by @lmacielvieira early.
Having CHNOPS based adducts will increase the chance for a generated decoy ion to be isomeric or isobaric to an ion of a real molecule. As this affects the FDR calculation, it's important to remove those adducts from the decoy selection pool.