I was recently using STAR to compare RNAseq data, and I noticed that it has a two pass mode with different way:
STAR can run in a 2-pass mode in two ways,
1) By re-generating the genome indices from splice junctions obtained
2) from 1-pass By using splice junctions directly during the mapping step
In the first method, we should run the STAR and generated splice junctions (SJ.out.tab), you can use these splice junctions for genome index re-generation, and run STAR with re-generated genome index again. But in latter method, we only need a params
According to the manual, there seems to be some advantage to using the former, especially when I have a lot of samples, but I've noticed that a lot of the literature seems to use the latter setup, why is that? How should this parameter be selected?