21Q1 Combined Achilles + Project Score CRISPR gene effect scores (linear modeling question)

Are the combined score data, scaled on common essential genes, suitable for linear modeling as a response variable?

The ANOVA models produced for Project Score [publication] identifying CRISPR effect genomic biomarkers were constructed on quantile-normalized, copy-number corrected fold-change values. I am trying to understand whether the gene scaling would create an issue.


Hi Ben,

The integrated data has been pretty extensively processed to remove batch effects, so a linear model should work fine. Of course no correction method is perfect. If you find anything weird, please let us know.

Thank you, @Joshua_Dempster .