search

jsh58 / Genrich

Detecting sites of genomic enrichment
MIT License
185 stars 27 forks source link

Genrich vs MACS2 or HMMRATTAC for ATAC-Seq #15

Closed jdidion closed 5 years ago

jdidion commented 5 years ago

I am looking to use Genrich for some ATAC-Seq analysis, but also see that you have a preprint describing an improved version of MACS2. Do you still recommend to use Genrich over MACS2? I also noticed Tao Liu has a new ATAC-specific peak caller (HMMRATAC) - do you have any opinions on that tool, or benchmarks vs Genrich? Thanks!

jsh58 commented 5 years ago

John,

Thanks for the question. I recommend Genrich for ATAC-seq, and this is reflected in the updated version of the ATAC-seq guidelines, in particular the section "Interpretation of alignments suitable for ATAC-seq", Figure 4, and the following paragraph.

The improvements made to MACS2 were to overcome some of its major bugs. They did not rectify its limited, inflexible alignment parsing.

I do not have any benchmarks of HMMRATAC. My opinion of it is low.

John Gaspar

dktanwar commented 5 years ago

I was about to post the same question and found that @jdidion already posted it yesterday 👍

I wonder if your benchmarking of Genrich vs MACS2 could be made available. Sorry, I tried to find for a paper, but, unable to locate any.

In the preprint of HMMRATAC, authors showed that it is better than MACS2

jsh58 commented 5 years ago

This is a work in progress. At this point, the best way to understand the differences between Genrich and MACS2 is by reading the ATAC-seq guidelines (as recommended above), and the description of the ATAC-seq mode of Genrich.