Bioinformatics protocols for differential RNAseq antisense expression

[antgomo]

HI,

I am analyzing data coming form Illumina directional RNAseq procedure, in order to find antisense expression. I ran Tophat with the corresponding flag, --fr-secondstrand, following the instructions from this previous thread

http://seqanswers.com/forums/showthread.php?t=9303

However is no defined how to deal with the bam files, I mean I got my two samples, and I don't know which algorithm is capable to deal with the XS:+/-. I read in the thread above tht htseq-counts from HTseq could help me to do with this, but people disagree if you must use --stranded=reverse or --stranded=no.

What I have done since now, in the typica RNAseq simple procedure is to load the bams int R using Shortreads and the use edgeR, DEseq or bayseq.

May I use HTSeq or cufflinks? And in which proper way to get the reads mapped to antisense strands?

Thanks in advance

[pbluescript]

HI,

I am analyzing data coming form Illumina directional RNAseq procedure, in order to find antisense expression. I ran Tophat with the corresponding flag, --fr-secondstrand, following the instructions from this previous thread

http://seqanswers.com/forums/showthread.php?t=9303

However is no defined how to deal with the bam files, I mean I got my two samples, and I don't know which algorithm is capable to deal with the XS:+/-. I read in the thread above tht htseq-counts from HTseq could help me to do with this, but people disagree if you must use --stranded=reverse or --stranded=no.

What I have done since now, in the typica RNAseq simple procedure is to load the bams int R using Shortreads and the use edgeR, DEseq or bayseq.

May I use HTSeq or cufflinks? And in which proper way to get the reads mapped to antisense strands?

Thanks in advance

You can use the bamtools filter option to specify a particular strand from the XS tag and get only reads that have that tag. Then you could use just the antisense reads for downstream quantification.