How Do Mappers Handle Chromosome Ends?

[nw328]

I am trying to figure out how BW-based mappers like BWA and Bowtie handle reads that map at the end of the reference and extend off the end. One example would be reads from the origin of a circular chromosome, where part maps to the beginning of the sequence used to build the index, and part of it maps to the very end. How are such reads scored? Is each extending base considered a mismatch, gap extension, or something else?

Thanks in advance!


NOTE:Also posted on Biostars

[tristan dubos]

Hi,
I was with the same problematic when I studies repeats region to determine number copy. I saw in BWA(mem) no mapping to the extremity of my reference (500 bases). To solve the problem by put 2 times my references aside. I tried before this to put N to the extremity of the references and low stringency alignment parameters but i was i low complexity sequences the i had everything mapping on the position that is why i used the first solution.
I hope it could help

Tristan

[maubp]

Very few mappers have support for circular genomes (e.g. CLC mapper), and worse, the SAM/BAM format does not handle this well. As suggested by Tristan, mapping to a doubled linear reference can be a useful alternative (but brings with it a lot of complications).

If you are mainly interested in mapping to the origin region of a circular chromosome, it would be simpler to make a linear chromosome split elsewhere in the circle, and map to that.