orb: DE wrong baseline #8
Description
Hi @tensulin, I don't know which repo stores the orb evaluation. Can you point me to it? We might want to move this issue to there?!
Looking at your confusion matrix, e.g. /vol/jlab/tlin/all_project/nf_results/refbasedassemblereval/oak/calculatedgeconfusiondeseq/deseq2_dbg_linear_cm.tsv, I find that the total number of genes is 47576+88+138+1034=48836. However, there are 5235 genes with zero counts in all samples combined. Does not sound to be fair as a baseline?!