broadinstitute / chronos

Modeling of time series data for CRISPR KO experiments
BSD 3-Clause "New" or "Revised" License
28 stars 9 forks source link

Likelihood analytical #15

Closed joshdempster closed 4 months ago

joshdempster commented 4 months ago

Changing to use the log ratio of the likelihood of the data (sum over replicates and guides per gene and cell line) as the metric for determining significant difference in gene effect between conditions. A lognormal model is used for the null distribution of likelihood ratios, with the null defined empirically by swapping condition labels as before. This allows p-values smaller than 1/len(null).