Talking Points Thursday - Putting the Complex back in Complex Diseases

So each Thursday, I thought I would blog about some current talk in genetic or anthropology or the combination of them both. As I blogged about the Havasupai Indians and the finalization of their case against Arizona State University, I thought genetics would be more appropro this week.

There is a recent article being disected on the blogoverse (or at least the one I frequent) that rehashes the argument (which is now getting old) regarding common vs. rare variants in GWAS studies. Other people have done a much better job of disecting this article and this post is incredibly informative. My argument is that everyone knows GWAS works to an extent but that it is still not identifying the missing heritabilty or the lack of variation described by significant SNPs. This is however because these SNPs represent only a small portion of the whole interaction occuring between genes that are involved in pathways. The genome is part of a biological system that is intergrated and inherently complex.

The basic idea regarding complex disease genetics is that the phenotypes involved are "complex"! This means that they are made of several genes that interact to create a protein and that they are also affected by the environment. A problem with candidate gene studies is that focus on variation within a single gene, without regard to the other genes or the biological pathways involved. So when a GWAS reports a gene to be involved in a complex disease it is only the tip of the iceberg. I'm not going to argue that GWAS is not informative, but it is an explatory statistical method. The idea proposed in the Mckellan and King article that the SNPs are not functional identfied in GWAS, totally misinterpets the method. The idea of GWAS is to identify regions of the genome that may be involved in a phenotype at higher resolution. GWAS works on the basis of linkage disequilbrium and provides 1 centimorgan region around where association occurs. This is ten-fold increase over linkage studies based on STRs. Besides, there are certain genes we know are involved in complex disease genetics because experimental work has been conducted on them in mice, rabbits, E. coli, etc., like hepatic lipase with HDL-C and variants with this gene show up in GWAS studies.

An alternative approach is to intergrate gene expression data and SNP data in order to identify the biological pathway involved in your trait of intrest. A recent paper in the American Journal of Human Genetics entitled "Integrating Pathway Analysis and Genetics of Gene Expression for Genome-Wide Studies" does a good job of describing this along with a certain type of methodology involved. One of the current problems associated with the joint type of analysis is that there is no set methodology and so different people analyze this in different ways. This paper represents a good start at this and may help in better identfying genetic variants that are impacting protein expression, which are then altering biological pathways that lead to chronic diseases. This is what will lead to a better understanding of the complexity of diseases, not denser chip set.