Match-between-runs and multiprotease search are not compatible

[rmillikin]

MBR attempts to map identifications from one run onto peaks in another run. This is incompatible with the multiprotease search because it does not make sense to map IDs from one protease onto another (e.g. gluC peptide IDs onto chymotrypsin peaks). There should probably be one FlashLFQ engine created per protease to match-between-runs between those files, and the results should be combined somehow (??).

[rmmiller22]

normalization -normalize within proteases will work but but protease data will not be normalzed between the proteases -makes sense because wouldn't compare GluC intensity to Trypsin Intensity

• probably ok
• identical sequences between different proteases should be ionized nearly identically (some risk of normalization not working properly there)

[rmmiller22]

Have experimental design that has bioteps, tech reps etc for each protease error checking must allow for this

[acesnik]

@rmillikin will your new protein quantification work with multiprotease data once this is resolved? Not that I have an application for that, but I wonder if anyone's ever done that and whether it improves quantification like it does parsimony.

[trishorts]

I would think that this would work fine. Fold change should be preserved irrespective of protease

[trishorts]

Normalization across bioreps at the whole protein level should not be effected by protease

[rmillikin]

Right now, normalization and MBR probably get messed up by multiprotease data. With Rachel's upcoming edit they should both work fine. Comparing protein intensities between a GluC and Trypsin file doesn't make sense for a variety of reasons but GluC vs GluC within a multiprotease dataset will work. Right now it would be buggy and normalization may crash.

[acesnik]

How's this coming along?

[acesnik]

@rmillikin has this been solved?

[rmillikin]

No - still a bug

[zrolfs]

This is the last issue in our 1.0 Milestones!