multiple mapped reads

[aquleaf]

Hi.

We did RNA paired-end sequencing. After alignment analysis, I found there were such reads. One end were with multiple alignments, while the other end with unique alignment. How to treat these reads? Should I eliminate them before inputting the alignments into Cufflinks or Scriputure?

Wish your suggestions! Thanks very much in advance?

[chadn737]

If one of the paired ends maps uniquely, then the other multi-mapped end would have came from that same transcript. I think you confidently trust that and discard the other regions that the one mapped to.

[cedance]

Yes, that seems to be the consensus. Although I discard this read with multiple hits and consider that unique hit as a single end read (because my library consists of paired end and single end read anyways and so I have to run tophat twice). But it seems perfectly reasonable to consider the solution from chadn737.