When Mike asked me what I wanted for my birthday this year, I had an answer for once. It just wasn’t one he was expecting. I wanted to get my genotype.
About ten years ago everyone got very excited when the sequencing of the human genome was announced. All three billion bases of DNA decoded -but, of course, not every base is the same in every person. The variation is what makes us individuals. Well – that and the course of our development and upbringing, in and out of the womb. Identical twins have identical genomes, but they’re not identical people.
The sequencing of the first human genome was a rather expensive endeavour. Advances in technology have brought the magic $1000 genome closer and closer, but we’re not there yet. In the meantime, there’s an alternative: genotyping. This focuses on very specific points on the genome (Single Nucleotide Polymorphisms, or “SNPs”) at which the particular base an individual has is linked to particular variants of nearby genes. For example, there’s a SNP nearby a gene controlling eye colour – one variant makes you overwhelmingly more likely to have blue rather than brown eyes, but it’s not the SNP controlling your eye colour, it’s that versions of the SNP tend to be inherited along with versions of the gene.
Companies like 23andMe will take a sample of your DNA and tell you what variants you have at about a million SNPs. This can tell you things about your health and genetic traits – if you have an increased or decreased risk for some diseases, if you’re a carrier for a genetic disease you may not know about, or just if you’ve got the gene for dry or wet earwax. (Presumably you know this already, but it’s cool to see if your genetics actually predicts your reality – most genes aren’t 100% predictive, just pretty close.)
It can also be used to see which stretches of DNA you share with other people, who are almost certainly – if those stretches are identical – some sort of relative. (Obviously you share more and more with closer relatives – it trails off very quickly after you get to about third cousins.)
It’s not as telling as you might expect. There are very few things, in all the variation of humanity, that we can tell for certain from your genes. Development and environment can affect some things as much as inherited traits. Other things are controlled by so many different genes that it’s hard to parse out from your genotype what your particular variants do; take height, which in the absence of environmental factors like malnutrition as a child is almost entirely controlled by your genetics, but not enough by any one gene (or small set) that you can point to it and say what a person’s height will be. On the other hand, there are a number of things that are controlled by a one or a small number of genes and possible variants of genes.
The actual testing kits aren’t that impressive – just a tube to spit in and a box to send it back in, really.
Mike and I duly spat into our tubes 23andMe (Mike on the grounds that “It would be weird if you got your genotype and I didn’t”) and posted them off. Then we waited.
Firstly, the results included the raw data – not just the nifty tools on the 23andMe site that tell you the percentage of your increased and decreased health risks or your percentage of Neanderthal DNA (I’m 2.6%, the Northern European average) – but the actual genotype data for every SNP, which is cool from a playing-with-bioinformatics-tools perspective. And stuff like finding out which variants I had at the SNPs from this paper (Stewart et al. 2010, oh yeah. Which is only interesting to, like, three other geneticists and me, but kind of a cool callback to that summer studentship and all the time I and my right thumb spent getting very friendly with the multi-channel pipette.)
All the predictive tools were kind of cool, too – my genes are correctly predictive of my eye colour, blood type, and that sort of thing, although laughably inaccurate in others. A good lesson in statistics and probability; when it tells me I have a “reduced chance” of having a photic reflex (that’s the thing when you sneeze in bright light) it reminds me that a “reduced chance” is a reduced number of people across a population of people with that genotype, because for me personally the chance is 100%. I have a massive photic reflex.
Health-wise the results were pretty boring, with one quirk: I’m a carrier for Tay-Sachs, a recessive genetic disease largely known for being prevalent among Ashkenazi (Eastern European) Jews, due to high rates of endogamy, a.k.a. inbreeding. This isn’t totally odd; Tay-Sachs is carried by 1 in 300 Northern Europeans, hardly world-beating odds. It’s associated with Ashkenazi Jews because their rates are more like 1 in 30. (They aren’t actually the only Northern European group this is true for, either; just the most well-known.)
But it also so happens that there’s a family story that my maternal grandmother’s paternal grandmother and her family were Russian Jews who emigrated to Australia. It’s pretty light on detail, but there’s nothing to suggest it’s wrong. The aforementioned endogamy makes it easier to spot even small amounts of ancestry from this group, because you’re far more likely to share identical stretches of DNA (or, in the case of genotyping, identical runs of SNPs) even with distant relatives.
As it happens, the “relative finder” on 23andMe, which checks to see if you share significant identical stretches of DNA with other users – who are almost certainly your distant cousins – brings up a few people with Eastern European and/or Ashkenazi ancestry. They’re small stretches, but still identical. I also ran my entire genotype through the EuroDNACalc tool, an R script devised to search for Ashkenazi Jewish ancestry in Northern Europeans, and it came back at 6%. This is near exactly the amount of DNA you’d expect me to share with a great-great-grandparent – i.e. my grandmother’s grandmother. The 95% confidence interval is 0-22%, so it isn’t conclusive, but it’s intriguingly congruent with the family story.
Right now I’m looking around for other bioinformatics tools I can use to analyse my raw data – I think playing with it is going to be an interesting hobby for some time to come. (Until those genome sequencing prices get to below $1000, though. Then the real fun happens.)