I am using WindowBed, part of the BedTools suite, to align reads to a reference file and I obtained a very interesting result. I am trying to rule out an analysis artifact that could be caused by extending the reads or by aligning read midpoints rather than 5' ends. It is my understanding that WindowBed aligns the 5' end of the read to the reference point, rather than extending than mapping the read midpoint, or extending the 3' end of the read and mapping the midpoint. Am I correct in this assumption, that the 5' end of the read is in fact what is being aligned?
Any help here would be appreciated. The BedTools manual, which is very good, doesn't seem to address this.
Thanks