Z-score normalization relative to negative control distribution

ArcInstitute / ScreenPro2

Flexible analysis of high-content CRISPR screening

https://screenpro2.rtfd.io

Other

21 stars 4 forks source link

Z-score normalization relative to negative control distribution #17

Closed abearab closed 9 months ago

abearab commented 11 months ago

root – https://github.com/GilbertLabUCSF/ScreenPro2/issues/26, also see https://github.com/GilbertLabUCSF/ScreenPro2/issues/22#issue-1989184854

abearab commented 9 months ago

#calculate epistasis interactions, optionally z-standardizing based on negative controls
emap1.loc[sgRNA,:] = epistasis / epistasis.loc[singlesTable['gene'] == 'non-targeting'].std()

source: https://github.com/bherken/herken2022_originalcode/blob/main/BWH_call_genetic_interactions.ipynb

abearab commented 9 months ago

I just performed a comparison between ScreenPro2 and ScreenProcessing results. I used the Decitabine CRISPRi-screen results for MOLM-13 cell line as a benchmark.

I think current implementation for phenotype calculation at v0.2.6 is very close to ScreenProcessing.