Bowtie2 mapping algorithm

Question

[Ribas]

Hi all,

Could you help me to understand the following problem? I am mapping reads to a viral reference genome, including different genotypes (someone very similar).

This is my actual strategy:

• First mapping allow me to find the best genome.
• Second mapping is against the choosen one, in order to increment the reads mapped.

This works, however sometime I get weird results.

For instance the total reads count is always greater but in some regions there are less reads!

How can be possible? Bowtie2 algorithm takes into account the reference length?

Many thanks, Diego

[h.mon]

You should include the commands you use, and illustrate your post with a particular example of the puzzling coverage issue. Did you check the mapping sam flags? Maybe it seems less reads because there are less supplementary / secondary mapping in some regions?

[Ribas]

I am running bowtie2 with --very-sensitive-local alignment. I do not think is regarding supplementary/secondary reads since keeping only primary one.