How do we balance the need for cost savings (stack more perts per sample) against the chance of killing the samples? I wrote up various assumptions here.
https://ekernf01.github.io/multiplexed_perturbation_studies/ (5/n)

Comments