Mapping position bias?

[tdyo]

Hi all,

This is Illumina HiSeq2000 paired-end sequencing mapped to a reference transcriptome using Bowtie2 (using the default end-to-end alignment approach, if it matters). I took the mapped positions from the SAM file and generated a histogram, and the mode is easily base 1 on the scaffolds. Is there any particular reason for this? Is it something to be concerned about? Thanks for any feedback.

[westerman]

Shouldn't you make some compensation for the size of the transcript you are mapping to as well as the overall PE length. If the transcripts are short then you would expect the start site to be the first base. Ditto if the PEs span a long distance.