Haplotype Determination With Next-Gen Data In Nuclear Families

Question

[Biomed]

I need an advice concerning determining haplotypes small pedigrees/nuclear families, using next gen sequencing (whole exome) data. I have whole exome data from families with mother, father and between one and in the best case 4 children. I also have Illumina infinium chip data on all of the individuals.
I am mainly interested in determining the haplotypes using next gen sequencing data but any robust method that may also integrate chip genotype data and whole exome data would be valuable as well. We have some experience with ENT but many other tools were mentioned before (i.e. http://biostar.stackexchange.com/questions/551/reconstruction-of-haplotypes-from-the-genotypes-on-snps ) Can you please advise on how to best approach this problem?

Thank you

[Samar]

Hi..

I wanted to perform similar analysis of finding linkage/haplotype information using exome seq data in families. Can you please share how you have finally worked it out? Are there specific softwares that can work with vcf files?

Any help will be highly appreciated.

Thanks

Samarpana

Answer 1

[Aaronquinlan]

Perhaps I am being naive, but wont haplotype construction from an exome capture experiment be rather limited? You are only capturing exons, which are typically on the order of 100-500bp, and most exons contain no polymorphism. To discern a haplotype for a single individual at a single exon, would you not need at least two heterozygotes and reads or pairs of reads that cover both sites? Otherwise, you'll be left without clear evidence of which alleles are on the same haplotype.

In other words, it is conceptually possible, but my feeling is that you only have power to do the directly from you capture data for a very small portion of the exome.

Please let me know if I am missing something.

[Biomed]

I agree with most of what you say but with approx. 100 exomes sequenced for our group we see a lot of variants in the exomes. You may argue that these are not true polymorphisms but a lot of those variants are known snps with minor allele frequencies that would qualify them as polymorphisms. I agree that it is conceptually possible to do it with exome data and the real deal is using whole genome data to do this. I am mainly interested if anyone thought about this or attempted to do anything similar to this. Thanks for your interest in the subject.

[Aaronquinlan]

interesting. what fraction of exons do you see having > 1 heterozygote? from my experience, it's some like 80% have 0 mutations, 10-15% have one, and so forth.

[User 1244]

I would use PHASE for haplotype reconstruction.

Answer 2

[dsmarcoantonio]

It seems that someone addressed this question. There is a new methodology using whole-exome to call haplotypes and look like promissor.

Accurate local-ancestry inference in exome-sequenced admixed individuals via off-target sequence reais. Hu, Youna; Willer, Cristen; Zhan, Xiaowei Kang; Hyun Min; Abecasis, Gonçalo R. American journal of human genetics. 2013