Paired end stranded wiggle/bedgraph output

[Jake Freimer]

Hi Alex,

I was just wondering how paired end stranded RNA-Seq data is handled when calculating coverage for wiggle/bedgraph files. Read1 will be on one strand and Read2 will be on another strand. However, looking at the wiggle files I only see coverage on the strand that is the same direction as the gene. Is STAR figuring out the coverage for both Read1 and Read2 together and putting it on the correct strand, only using one read, or doing something else?

Thanks

[Alexander Dobin]

Hi Jake,

with --outWigStrand Stranded, STAR will revert the strand of the 2nd read, and create two wiggles, str1 and str2. 

str1 will contain reads where read 1 maps to +strand, while str2 - those with read1 mapping to -strand.

Depending on the strandedness of your protocol, you will need assign actual RNA strands to these files, e.g. for Illumina Tru-seq, str2 is +strand and str1 is -strand.

Or you can just check which orientation follows the annotations.

Similar to GeneCounts, STAR does not decide what is the strandedness of your library, it simply outputs both options.

Cheers

Alex