Calling Of Rare Variant From The Illumina Human Exome Array

Question

[Thomas]

Dear all

Our group has started chip genotyping our study samples with the Illumina Human exome array.

On this chip array, a big proportion of the SNPs are very rare and my concern is that it will be very difficult to call true heterozygous genotypes among the many homozygous.

My question is: Does anyone have experience with the Illumina Human exome chip and can you provide ideas for calling these rare variants.

Thanks for any comments

Best wishes Thomas

[Alex Paciorkowski]

@Thomas -- did you also genotype parental samples? This is one way to help you sort out homozygous vs heterozygous. (ie homozygous calls have to come from somewhere -- I would then expect the parents to be hets)

Unfortunately, I don't know much about the qc and coverage metrics for this array. But there may be features there that can help you sort out good data from bad.

[Thomas]

great and thanks for the comments. Unfortunately we do not have family data.

Answer 1

[hansa]

I have some experience with Illuminas HumanExome chip and generally the chip perform very well compared to population frequencies reported from the 12000 exome sequences. However a small number of variants give distinctly wrong answers, but many of these are obvious upon direct inspection of clusters. We have only run cases and our main problem is to find suitable control frequencies that are determined by chip (not exome sequence). If anybody knows where to obtain exome chip genotype I'd love to hear about it.

[Thomas]

Hi Hansa, I think we should talk together. We have around 5K population-based controls (Europeans). Allthough I need to talk with my PI to confirm I believe we are very open and keen for collaboration. I would also very much like to hear how you are handling the data, since I do not have that much experience. Please email me if you are interested ([e-mail]).