Right now, it is computed with the pingouin package (partial_corr) but the conditioning on the cell line and drug is done by factorizing the identity. This is wrong, we need a continuous measure where it is okay to interpret it as number. See EFFECT suite for inspiration.
Right now, it is computed with the pingouin package (partial_corr) but the conditioning on the cell line and drug is done by factorizing the identity. This is wrong, we need a continuous measure where it is okay to interpret it as number. See EFFECT suite for inspiration.