search

Noble-Lab / casanovo

De Novo Mass Spectrometry Peptide Sequencing with a Transformer Model
https://casanovo.readthedocs.io
Apache License 2.0
112 stars 40 forks source link

Add min_peaks config option #264

Open wsnoble opened 12 months ago

wsnoble commented 12 months ago

Casanovo should have a configuration option that specifies the minimum number of peaks that are required to be in a spectrum before it will make a prediction. Currently, Casanovo will happily make a prediction (sometimes with a high score) even if the spectrum has only a handful of peaks. The default value for the analogous parameter in Tide is 20 and in Comet is 10. For de novo, I think we should use a default of 20. This is the minimum in the 9-species benchmark.

Note that, ideally, Casanovo would track and then report something like "Info: XXX spectra skipped because they have fewer than XX peaks."

bittremieux commented 10 months ago

Postponed until functionality from DepthCharge v0.3 has been integrated, which will make this very easy to implement.