Archiver > GENEALOGY-DNA > 2008-06 > 1212695338

From: Didier VERNADE <>
Subject: Re: [DNA] SNPs on chip
Date: Thu, 5 Jun 2008 21:48:58 +0200 (CEST)

Ann Turner wrote :

> I'm away from home right now, and my Internet access is subpar, so I haven't
> located any detailed explanation of the BeadChip technology with diagrams etc.
> Roughly, they create miniscule beads with short DNA fragments attached. The
> fragments will include the alternative versions created by known SNPs. The chip
> can include many hundreds of thousands of such fragments. Then your DNA is
> chopped up into little pieces and floated over the beads. If your DNA is the
> complement of one of the fragments, it will hybridize with it. There are
> fluorescent dyes that light up when that happens, and the light patterns are read and
> interpreted by automated software.
> The design process for a chip takes quite a bit of overhead investment, not
> to mention the manufacturing costs for each unit. There is a chip from
> Affymetrix for the complete mtDNA molecule. There should be an article out shortly
> describing how well this works, as a number of sequences were deposited to
> GenBank recently. Various technical articles have already used this chip, and while
> the accuracy might be sufficient for group studies, it strikes me as being
> less than an individual would desire.

Thanks for this good summary of how this chip technology works. As there are over 500 000 SNPs detected by a chip I am not sure how they find different wave length for each pair of DNA on the beads (for each SNP there is one DNA sequence and the corresponding DNA with the derived state, as I understand) . Still the explanation as it is really helps me to understand the limits of this technology so thanks again Ann.

By the way 23andME provided me with the large (over 5 Mo) file with "all" the results. I didn't download it and I asked them if I can get a separate file for Y and mitochondrial SNPs. If you don't mind I would prefer to communicate only on this reduced set of SNPs. I am now waiting for their answer. I put quotes on "all" because it seems I got some "no call" results and i am also asking them to get results (ancestral or derived) for these. I'll keep you informed.


This thread: