Bowtie index size / richness matters ?
1
0
Entering edit mode
7 weeks ago
maschi88 • 0

Hello,

I have some problems with bowtie these days and I don't get it.

I am working on smallRNA-seq, to identify different small RNA species, notably miRNAs. My animal organism is not a model species so I don't have any reference available for it. I decided to use 3 close species - that are referenced - to identify my reads. For miRNAs, which are well conserved, it is a known practice that works well.

So I built my miRNA index with a multifasta, containing all the references from the 3 species, and surprinsingly Bowtie doesn't find a lot of miRNA reads, it is like 12% of success, even with mismatches.

When I go through the unaligned reads, I am intrigued by a sequence that is very abundant (like 7M copies) and with a size that perfectly fit miRNA. So I decided to do a blastn with it and it gives me a genome region that it is known to contain a primary miRNA (which produces mature miRNAs), it is highly conserved through mammals, so this read must be a miRNA.

Going back to my references, I checked if this known primary miRNA is present... and it is, for the 3 species. Meaning I have at least 3 references, with no variation, containing the sequence of my overrepresented read. But Bowtie failed for any reason.

By curiosity, I tried to map my library again, but using only one species at a time for my index and this time Bowtie maps the reads. Giving me like 60% of miRNAs.

How exactly Bowtie's index works ? It seems like the more it has to use, the less it maps.

I used the default parameter to build my indexes. And the same mapping condition in my different tries.

Thank in advance.

bowtie • 488 views
ADD COMMENT
1
Entering edit mode
7 weeks ago

It is a read multimapping problem - bowtie1 (?) likely considers these repetitive and filters them out. Your mistake was to combine all 3 references and create one index. Create 3 separate bowtie indices and align to each sequentially if you like.

edit - Sorry - I see, you did this later. Read up on read multimapping on the internet to see all the great answers to this quite elementary issue.

ADD COMMENT
0
Entering edit mode

I see, thank you. I'll change my code and build several indexes then.

ADD REPLY

Login before adding your answer.

Traffic: 1179 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6