Test for recombination in a clonal tetraploid?

Question

[Adrian Pelin]

Hello,

I have been searching this subject and trying to find a technique that fits my needs and so far I have been having no success.

I am looking to test try and find crossover events in my HiSeq datasets. I have 9 strains of the same non-model clonal species that are very similar to each other in that sense that a lot of genomic variation due to ploidy is shared. I am now interested if despite clonal propagation, there is still recombination occurring.

All I have is 9 HiSeq 100bp x 2 PE reads with a fragment size of ~300bp for my 9 strains.

Any recommendations?

Thank you,
Adrian

Answer 1

[Ying W]

I am not an expert on this subject but since nobody has replied, this would be my guess at how you would start:

1. Pool samples from your HiSeq runs, de novo assemble to create reference (using tools like velvet
2. Align reads from sample to newly created reference
3. Perform SNP calling across samples and look for blocks of SNPs inherited together

Maybe you can use these blocks of shared SNPs to infer if recombination has occurred. A quick search turns up this paper on assembly and recombination rate estimation of a non-model organism that might be useful.

[Adrian Pelin]

What would you describe as blocks of SNPs inherited together? I have already done steps 1-3 except I did not pool them together out of fear of chimeric contigs.

The paper seems to talk about SNP arrays, and I am not sure how that links to what I am trying to do.

[Ying W]

When the crossover event occurs, a block of SNPs would be on a chromosome in an order that is different from usual. Thinking about this some more, I think you would need to try and infer what the genotype of each chromosome is and then from there figure out if there is a crossover event. Here is a simple example to illustrate this

Strain1
n=1 ----T---C---A---A--
n=2 ----T---C---A---A--
n=3 ----A---G---C---T--
n=4 ----A---G---C---T--
Strain2
n=1 ----T---G---C---T--
n=2 ----T---C---A---A--
n=3 ----A---C---A---A--
n=4 ----A---G---C---T--
Strain3
n=1 ----T---C---A---A--
n=2 ----T---C---C---T--
n=3 ----A---G---A---A--
n=4 ----A---G---C---T--

From this, assume that strain1 is your reference, strain 2 would suggest a crossover event between n=1 and n=3 somewhere between the 1st and 2nd SNP while strain 3 would suggest a crossover event between n=2 and n=3 between the 2nd and 3rd SNP. This is the kind of stuff you are looking for right? If so, you would first have to figure out which SNPs come from which n=# and to do that, you would have to look at your reads and see if the same SNPs exist on the same reads and go from there. I don't know if there is any software that would do this for you but if you do end up thinking of a good example to show what you are looking for, I would re-post this question with a more general title and explain the tetraploid / crossover stuff in your question.