View Single Post
Old 04-02-2014, 07:56 AM   #4
Senior Member
Location: USA, Midwest

Join Date: May 2008
Posts: 1,178

Originally Posted by areyes View Post
...For example, for a gene that is in the forward strand, you will get all the reads mapping to the reverse strand. This is a [b]it confusing!

So, what I recommend is that in your IGV browser you could colour your reads depending on the strand where they map to, and then see if what I describe above is the case. If so, I think the parameter "-s reverse" parameter of the script should do the trick.

More accurately, for paired end reads, read 1 will map to the anti-sense strand while read 2 maps to the sense strand, so in this case you will get reads mapping to both strands. You need to distinguish both mapping strand and read (1 or 2). The attached image from IGV shows mapping of TruSeq Stranded, paired end reads. In this case the gene is transcribed from the forward strand of the reference (left to right). Reads are colored by strand, blue == minus, red == plus (in this highly zoomed in view the reads glyphs also have arrows indicating their direction). The alignments are then grouped by "first-in-pair". The top group is read 1 and the bottom group read 2.

Alejandro is correct that for TruSeq Stranded Kit libraries you need to specify "-s reverse" for or htseq-count if you are counting reads to genes for DESeq(2)
Attached Images
File Type: png igv_panel.png (18.1 KB, 21 views)
kmcarr is offline   Reply With Quote