GenomeWeb Daily News points to a new funding opportunity from the US National Institutes of Health (NIH) for researchers interested in studying the link between genetic variants and variation in the expression levels of genes.
This is an incredibly important area of research. Genome-wide association studies have recently uncovered vast numbers of DNA regions linked to common diseases (the latest estimate I've heard suggests around 400 common genetic variants associated with 75 different diseases or traits) - but we still don't have a clue how the majority of these regions actually alter disease risk.
However, it's becoming increasingly clear that very few of these association signals are due to changes in the sequence of the proteins produced by genes. Instead, it seems likely that many of the underlying genetic variants affect human health through changes in gene expression - basically, by alterations in the levels of messenger RNA produced by a gene, with downstream effects on the amount of the protein produced by that gene. Proteins, by and large, are the molecules that actually do stuff (like holding cells together, or driving the chemical reactions required for life), so such changes can have profound effects on human health.
The link between genetic variation and gene expression has thus been a target of intensive research over the last few years, a fair chunk of which has been performed by colleagues at the Sanger Institute. The major limitation of this research is that it has almost exclusively been performed using a single type of cell - immortalised lymphoblasts, blood-derived cells with the handy property of being able to grow and divide in a Petri dish essentially forever.
Although this research has already answered fundamental questions about the link between genetic variants and gene expression, it provides insight into just one of the hundreds of cell types dwelling within the human body. Obtaining a broader picture of the way that genetic variation alters disease risk will require performing this analysis in many different tissues in hundreds of humans, all of whom have very detailed information on patterns of genetic variation.
That's the aim of the NIH's Genotype-Tissue Expression (GTEx) project, the focus of the current funding boost. GTEx is a pilot project that will look at the correlation between genetic variation and gene expression in many different tissues, collected during autopsy or organ transplant procedures from 160 donors. In other words, in addition to lymphoblasts, researchers will have access to data from tissues like liver, heart and brain that are typically off-limits - and that in turn will provide insight into diseases that primarily affect these tissues, information that simply couldn't be gleaned from studies of cultured lymphoblasts.
The NIH funding is aimed at supporting the development of new methods for detecting links between DNA-level variation and gene expression. This is extremely timely: new sequencing technologies offer images of astonishing resolution of patterns of gene expression, but techniques for making full use of these data are still in their infancy. A targeted cash infusion is a very handy incentive to drive these techniques forward.
Instead, it seems likely that many of the underlying genetic variants affect human health through changes in gene expression
Changes or differences? If it's changes, then using tissue after the disease has progressed doesn't seem sensible to me: you don't know what else has happened in between.
I'll also pimp a short commentary of mine where I pointed out that gene expression studies are useless at telling us what's really going on, because they say little about the underlying physiology.
I guess I meant "differences", i.e. people with different genotypes have different levels of gene expression.
I agree with your point about the dangers using tissue late in disease progression - by this stage there's every chance that the primary gene expression differences have been obscured by secondary changes (e.g. fibrosis, inflammation, etc.). However, most of the studies here aren't actually looking at disease patients: they're looking at genetic control of gene expression across a large cohort of healthy (or at least unselected) individuals. This doesn't necessarily provide deep insight into highly disease-specific processes, but it does allow a dissection of the effects of common disease-associated variants on expression, as well as some intriguing insights into the generalities (here's two examples).
Thanks for the link to your paper - I especially enjoyed the carefully understated first paragraph! I agree with your argument regarding the likely irrelevance of most gene expression changes to disease; given the sheer number of common eQTLs that have been discovered recently, I think it's likely that you're correct that the majority of these have essentially no effect on overall phenotype.
However, that's not to say that all of them are neutral, and I'd still argue that it's likely that a significant proportion of disease risk variants (possibly the majority) are the result of either direct or indirect effects on gene expression.