Decoy proteoforms should have decoy sequences

smith-chem-wisc / ProteoformSuite

Construction, quantification, and visualization of proteoform families

https://smith-chem-wisc.github.io/ProteoformSuite/

GNU General Public License v3.0

12 stars 19 forks source link

Decoy proteoforms should have decoy sequences #461

Closed leahvschaffer closed 6 years ago

leahvschaffer commented 7 years ago

Currently, decoy proteoform sequences being set to corresponding theoretical's actual protein sequence

leahvschaffer commented 7 years ago

***need to unparallelize shuffled_proteins loop for taking hunks off of giantprotein

acesnik commented 7 years ago

Why does it need to be unparallelized? Isn't substring thread-safe?

leahvschaffer commented 7 years ago

If you want the decoy database to be the same each time with the seed

acesnik commented 7 years ago

I don't see it; the Random object used for shuffling before that loop is seeded. Isn't the only issue here setting the decoy TheoreticalProteoform objects to have the correct sequence, so that the heuristics using that sequence are correct.

acesnik commented 7 years ago

Totally on board! Thanks for explaining. prevLength isn't thread-safe!

leahvschaffer commented 6 years ago

Fixed in #424