search

mckennalab / FlashFry

FlashFry: The rapid CRISPR target site characterization tool
Other
63 stars 10 forks source link

output all off-target score #18

Closed ryys1122 closed 3 years ago

ryys1122 commented 4 years ago

how to output off-target score for every single off-target , but not only DoenchCFD_maxOT and DoenchCFD_specificityscore of all off-target?

aaronmck commented 3 years ago

yeah, that seems possible though it may take a bit of rework (I think the Doench CFD score is the only that is meaningful for a single off-target).

ryys1122 commented 3 years ago

i think the program haved calculated Doench CFD score for every single off-target befor outputing DoenchCFD_maxOT, so what the score class need to motify is just including Doench CFD score in the offTargets column for every single off-target with score option --includeOTs, like this AAAAATAGGGAAAAGGGCACCGG_1_4_Doench_CFD_score<2:838572^R>,ACAAATATAGCACATGCCACTGG_1_4Doench_CFD_score<5:220383333^F>

aaronmck commented 3 years ago

I've put up a pre-release version 1.12a that supports this. During both the discover and score modules you'll need to have "--includeOTs" added to your command line. We have to support any number of these associated scores, so the output looks a little different than you proposed:

ACTGTCCCAGCAGCAGAAGAGGG_1_4<chr22:48437007^R>{Doench2016CFDScore=0.06444234483586098},

There are some more things to test, but I thought you might want to try it out.

ryys1122 commented 3 years ago

thank you very much, i'll have a try.

maximilianh commented 3 years ago

Sorry, just lurking here and listening in, and wondering why we still don't have a better off-target score than John Doench's CFD... isn't that weird? Maybe it's just too hard as a problem...

aaronmck commented 3 years ago

Waiting for the meta-analysis in CRISPOR2 to tell us a better metric :-)