Batch correction in Post-Chronos datasets

Q&A
1

Dear DepMap Team,

I am curious about the batch correction method used by DepMap on their post-Chronos datasets, given the availability of multiple batch correction pipelines as mentioned in ‘Integrated cross-study datasets of genetic dependencies in cancer’.

I am also interested in performing a spheroid screening in a cell line, and would like to compare my own post-Chronos dataset with post-Chronos dataset in DepMap. However, our time points are different. Is it still possible to make meaningful comparisons between the datasets after batch correction?

Thank you very much,

Nicole

2

Hi Nicole,

The larger concern is usually library batch effects. If you’re doing a screen in a library other than Avana/Humagne/KY, there will be large reagent-based batch effects. The time batch effect is not too large unless your timepoint is very different (<14 days or > 26 days).

Either way, with only one model I would not attempt per-gene batch correction. There is no way to distinguish a batch effect from a genuine biological difference. What you could do that’s relatively simple is a quantile normalization, so at least the overall distribution of effects is the same in your model and DepMap.