chung-lab/SCAFE

pooling of multiple samples

Closed this issue · 1 comments

baigal628 commented

Hi,

I am curious about how the scafe.workflow.sc.pool is functioning to merge all the tCREs across samples. As mentioned in README, I ran workflow.sc.solo on an individual sample first and used scafe.workflow.sc.pool to pool all the 64 samples together. In the single sample run, I got ~15,000 tCREs on average. But after pooling only ~9,800 was returned as in common. I am thinking if the merging criteria are too strict for my dataset.

I didn't find a description of how scafe.workflow.sc.pool is doing the pool step. It would be great if you can help to explain.

Thanks!

jenchien commented

Hi, as far as I know, scafe.workflow.sc.pool pools all samples at the ctss level, then calls and filters tssClusters by the same workflow as in scafe.workflow.sc.solo. Therefore, the unfiltered tssClusters for each sample will not necessarily appear in the pooled output. How many tCREs do you get after the pooled run? When you say "after pooling only ~9,800 was returned as in common", do you mean on average there are about 9800 pooled tCREs intersecting tCREs in each sample?