Improving manual mass shifting performance

smith-chem-wisc / ProteoformSuite

Construction, quantification, and visualization of proteoform families

GNU General Public License v3.0

12 stars 19 forks source link

Closed acesnik closed 6 years ago

acesnik commented 7 years ago

Taken from Lollipop.regroup_components:

Could be improved. Used for manual mass shifting.

Idea 1: Start with Components -- have them find the most intense nearby component. Then, go through and correct edge cases that aren't correct.
Idea 2: Use the assumption that proteoforms distant to the manual shift will not regroup.
Idea 2.1: Put the shifted proteoforms, plus some range from the min and max masses in there, and reaggregate the components with the aggregate_proteoforms algorithm.

leahvschaffer commented 7 years ago

For identified families, since you added identifying experimentals, we could automate mass shifting in EE for identified experimentals

acesnik commented 7 years ago

True! I might try to keep the theoretical references in a linked list, then, so we can tell how far away it is from the database reference.

acesnik commented 6 years ago

Performance was improved by parallelizing aggregation, in particular.