Finding deferentially active regions with DESeq2

Question

[m.wekking]

I have a File that contains certain enhancer regions per chromosome and per region how many reads are found from different samples. The 1 and the 2 stand for the two different groups I have. The original file is tab delimited. I.E.:

$ cat enhancers.bed

chr | start | end | sample1a | sample2a | sample2a | sample2b
  2 | 2000 | 4000 | 3245 | 2345 | 4234 | 5409
  2 | 2200 | 4400 | 1842 | 2086 | 1902 | 2181
  2 | 3500 | 5200 | 6912 | 6209 | 4234 | 5409
 ...

Now I wanted to find in which regions a certain group is deferentially active. I read somewhere that I should use DESeq2 for this, now the only problem is that I cant find anywhere how I should do this ... I haven't used R in a while and my skills in this language are very rusty... Is there someone who can help me?

[ATpoint]

Read the DESeq2 manual, it convers everything you need. Alternatives for ChIP-seq are DiffBind or csaw, both are Bioconductor packages and both have extensive documentation.

Answer 1

[shawn.w.foley]

The DESeq2 vignette is very detailed and user friendly, I used it when I was just starting to learn R. Take some time to go through it and to model your experiment after their most appropriate examples:

http://bioconductor.org/packages/devel/bioc/vignettes/DESeq2/inst/doc/DESeq2.html

If this is ChIP-seq data then you might want to consider some of the equivalent packages that have been optimized for ChIP-seq rather than RNA-seq, such as DiffBind and DiffReps. There are a lot of differential binding algorithms that all give you distinct results. From experience, the same dataset can show drastically different results depending on the algorithm you use, so you'll need to think about what are some reasonable positive controls for your experiment, and determine which pipeline produces the most appropriate results. A nice review of many different algorithms is here:

https://academic.oup.com/bib/article/17/6/953/2453197