Extraction of (two) sequence variants from exons

Question

[tomasfer]

Hi,

I mapped reads (150 bp) to exon reference (in Geneious) and it is apparent that there are two different types of sequences, probably paralogues, each with high coverage

Making consensus does not make sense in this case, sequences will be used for phylogeny and paralogs can destroy true signal etc.

Is there any possibility how to extract these two sequences/paralogues from mapped reads, i.e. make an alignment (actually two alignments...) from these reads belonging only to particular variant.

I prefer some possibility that allows for streaming as I have several thousands of such loci (of course not all of them have two or more "paralogues").

Thanks for some suggestions or links.

Tomas

[Nicolas Rosewick]

Did you align your sequences against the whole genome? Or only on this specific exon sequence?

[tomasfer]

My reads are from enriched library (ca. 2,500 exons, non-model species) so I mapped all reads to "pseudoreference" consisting from sequences used for enrichment separated by 400 Ns (to avoid reads to be mapped to multiple exons). What do you see on a picture is a part of a single exon.

[Devon Ryan]

Even with enriched libraries, you're highly advised to align to the entire genome. That alone may clear this up.

[tomasfer]

OK, but as I am working with non-model taxon, genome is not available :-(

[Devon Ryan]

Here, "genome" is whatever sort of reference you have, be it just contigs or something else.

Answer 1

[Devon Ryan]

Assuming this isn't some sort of biased mapping (see NicoBxl's comment), the simplest route would be to call variants and then phase them. You should then be able to use the results to create two different reference sequences matching the two haplotypes (you might have to write something to do this, I don't know of a pre-existing tool). Alternatively, you could just process the phased VCF files directly for determining the phylogeny.