RNA-seq stranded protocol

[dfernan]

Hi,

I have RNA-seq data that was generated with Illumina TruSeq stranded protocol, according to which, as far as I understand, first mate reads are supposed to map to the same strand of the transcript and second mate reads are supposed to map to the opposite strand of the transcript.

Assuming this is correct and eliminating any read-pair mappings that do not obey this condition I get very few read mappings. I noticed that I have a lot of read-pair mappings where the opposite to my condition is true. I.e., the left mate maps to the opposite strand of the transcript and the right mate maps to the same strand of the transcript. Are these read-pair mappings reliable? Should I not discard them?

[lindenb]

cross posted: https://www.biostars.org/p/105598/