Decrease stringency of bowtie2

Question

[Wilber0x]

I want to decrease the stringency of bowtie2 when mapping. How can I increase the overall alignment rate?

My index is from one of two closely related organisms (species A). I want to find the sweet spot where my reads from species A align to the index at a very high rate, but reads from species B do not.

Which parameters should I alter?

[Friederike]

if you have a reference genome for species B, it's probably better to create a hybrid index using both genomes (the chromosome names should reflect the individual species name) and align to that.

EDIT: If you need more justifications/pointers from the literature for this problem, look for publications on sequencing of Patient-derived xenograft (PDX) and CTC-derived explant (CDX) models, which discuss the technical problems of aligning reads from two closely related species (human, mouse).

Answer 1

[h.mon]

Your stated goal:

I want to find the sweet spot where my reads from species A align to the index at a very high rate, but reads from species B do not

do not match the method you want to use:

I want to decrease the stringency of bowtie2 when mapping. How can I increase the overall alignment rate?

If you want reads from the "correct" species to map, but reads from the "incorrect" species to not map, you want to increase alignment stringency. For that, an easy way is to use the --end-to-end --very-fast settings.

Alternatively, there is a plethora of parameters to adjust, such as -D, -R (Effort options), -N, -L, -i and other options (Alignment options), and several alignment scoring options, such as --score-min (see Scoring options and Setting function options). The settings found for one pair of species wouldn't necessarily work for other species, and if the species are very close, it may be impossible to separate a large proportion of reads.

Do you have a reference genome for species B?