On Disease Association

Hello again, my loyal blogventurer. The hectic nature of my life continues, though to my joy genomes are starting to fall into place (computationally, that is, not by insemination. Though doubtless that has also been going on. If not, god help us all). However, I think all this business is leading to general mental decline. Yesterday, in a fit of flustered Britishness, I thanked an ATM after using it, an event which, on retelling. triggering a ‘not-adapted-to-the-modern-world five’ from a quick-witted colleague.

Rambling introductory material aside, a few things have occured in the past week that have rekindled my passion for medical genetics. A few of us put together a talk on monoclonal antibodies (an awesome medical technology, which if you are lucky I may talk about in the future), and I signed up to work on disease gene association in the coming months. The latter I have a special affection for, an affection which I shall now point at you.

Genes for Disease

Gene association is the name given to the attempt to find correlations between gene variants and other traits. The modern flavour of this act is Genome-Wide Association (GWA), which involves scanning the entire genome looking for changes in the genome that are associated with diseases. The largest of these studies is the Wellcome Trust Case Control Consortium, which compares the genomes of healthy people with those suffering from complex diseases such as diabetes, Crohn’s Disease or depression. In essence, they count millions of variant genes, and look for ones that occur more in ill people than in healthy ones.

So why do we care about this information? The most obvious one is so that in the future we can genotype individuals, and tell them what diseases they are at risk for. Often this information will be useless, but it can sometimes be used to make lifestyle decisions. For instance, mice can have a gene that makes their inner-ear cells susceptible to damage; if a human had that gene, they would do well to buy over-ear, noise canceling headphones, rather than in-ear ones, to avoid hearing loss in later life.

More usefully, gene association data can be used as a starting point for research into how the disease comes about in the first place. Regardless of whether a disease is caused, in part or in whole, by a gene variation, it will manifest in the form of a change in the biochemical makeup of the body (depression manifests as changes in brain biochemistry, regardless of the degree to which it was caused by them). Studying how the onset of disease interferes with body function is very difficult; it is often hard to know where to start. However, if you find a gene that has some effect on the disease, even if it is in a tiny minority of disease cases, or only has a very small effect, it tells us that the gene products (proteins or RNAs) of this gene have some role in the disease. A good example is people who lack certain T-cell receptors, and seem to get HIV infections far more rarely. We can use these molecules as starting points, places to begin designing experiments to give more in depth explanations, hopefully leading to treatments.

Exciting Complications

One thing I like about GWAs is that they force you to think about all the things that can effect disease risk, by throwing up little design problems. For instance, to stop yourself getting odd answers you need to account for population stratification; you can bias your answers by over-selecting certain people in your case or control samples. If all your Diabetics come from York, and all your healthy people come from Surrey, you are going to end up concluding that Viking genes contribute to insulin resistance.

As in all things in this world, the answers you get from gene association studies can be dependent on the questions you ask. For instance, the WTCCC found a that variations in the gene FTO were significantly associated with the risk of Type II Diabetes; however when the Broad Institute tried a similar thing, they did not find any link. The reason for is that the FTO gene (which, as far as I can tell, stands for “FaT Obese gene”, though perhaps it is just trying to capture a general feeling of being Rotund) is a gene that has been shown to be associated with obesity. The WTCCC, which compares randomly selected individuals, found this gene tended to make people overweight, which tended to trigger diabetes, however the Broad Institute only compared people with similar BMI’s, and thus did not find genes that causes diabetes via obesity. Both answers are correct, in their own way, but neither fully captures everything that is going on.

Gene association can also be complicated by culture. A recent study showed that being of South Asian descent in Scotland was significantly correlated with Type II Diabetes , which cannot be account for by a difference in quality of care. Genetic link, I hear you chime with child-like glee? Perhaps. Though if the Indian’s I know are anything to go by, being South Asian is also highly correlated with having a number of elderly female relatives who produce, with supernatural efficiency, great amounts of almond sweets and fatty curries, and then smile insistently at you until you have eaten them all. I’m sure you can readily accept the proposition that this contributes to insulin resistance. A distinctly less pleasant example is that many African-specific neutral alleles in American cities will be associated with death by gun crime, through this would very much not be a genetic disease.

Finally

I have never really had the required medical knowledge, and, between you and me, the patience, to do real nitty-gritty medical research. I have a lot of respect for people who spend their lives tracking down exactly how cancer multiplies or HIV infects, and developing distressingly fiddly little molecules to inhibit them; I wouldn’t know where to start. However, as a geneticist I like the idea that I can sit further back, using statistical inferences to fish out little clues about the action of disease that can contribute downstream somewhere. Especially fulfilling is that I know that, as a Wellcome Trust researcher, everything I do will be available free on the web for biomedics In The Field to use in their own research. This is the essence of the Science 2.0 thing; different people and different methods flowing freely around the community, to be used to do real good.