Mapping to related species

[Guest]

How much mismatch is enough and not too much?

If I want to map my reads to related genera/species (within a family) how much mismatch should I allow. Obviously, if I set no or very little mismatches I will find conserved regions among them.

The question is - how to inspect how much their genomes are similar by just mapping reads? Should I just stick to stringent criteria and just go the way -> the more similar they are, the more reads will map even with no or little mismatches?

I will do whole genome alignments, but until I get a good assembly I thought I could try it this way. No?

Thanks!

[severin]

which species?

The other important question is whether there has been a lot of whole genome duplication or a lot of paralogues in the genes you are interested in looking at.
Many plants have a lot of more "recent" whole genome duplications than animals for instance.

You could always use a loose constraint say 30% mismatch and see what aligns. You could also plot the number of reads given a certain level of mismatch. You can filter the alignments after they are aligned based on whatever cutoff makes the most sense.

[lorendarith]

ATM I'm not really interested in any particular genes, but am doing this in order to get more insight on the relatedness to other species with sequenced genomes.

The problem is that the genus of the species I'm working on has no distinct placement in the family tree of e.g. Brassicaceae. Based on phylogenies with different markers, the closest related species with a sequenced genome also differs from publication to publication.

I guess what I really want to achieve with this is to preform some sort of synteny mapping without large pesudomolecules.

[MicroBio]

Going off of what severin was saying, do you think you could use a loose mapping constraint to pick out some regions of interest (areas where your reads map to other species), and then perform a PhastCons type of analysis to see how much conservation there is in that region between your genome of interest and those with related sequences? This way you would pick up weakly matching regions (because of the loose mapping constraint) that would show low conservation between genomes, and high matching regions with higher conservation scores.