Here's the problem with assuming that all (or most) diseases are caused exclusively by rare variants of moderate to large effect. While it's true that these variants would be unlikely to be identified efficiently in a GWAS study, they would be highly likely to be pinpointed in a well-designed multi-level linkage study with medical resequencing. These types of studies have been around for ages. One of the major reasons why GWAS became so popular was because it was a new way to tackle some of the diseases that had resisted the linkage approach in the first place, which argues against a single moderate-to-large effect mutation causing the disease. This does, of course, assume that you have a decent sized family in which to study the disease, but many of these diseases have sufficient family cohorts to study, particularly when you add in medical resequencing as is being done now. I find it very strange that people are so married to the idea that genetic architecture has to be either "common SNP" or "rare variant". Why can't we have some diseases of each type, and some diseases with a combination of these architectures for the very same disease? I expect that when everything shakes out, that is exactly what we're going to have...common SNPs creating phenotypes that predispose individuals to have particular phenotypes (in particular environments), with the occasional strong effect from a rare variant rising to the forefront. And to address your argument about selection, that was a non-starter...selection isn't something that works in just one direction or at just one strength. It's a force that has variable directionality and variable force. If you haven't become familiar with Wright's shifting balance theory, you should.

I was unhappily surprised at the lack of nuance in Miller's article. I hadn't read it before, and it read like is a devotee of David Goldstein and hasn't interacted professionally with researchers with a more thorough grounding in quantitative genetics and population genetics (Eric Lander, Leonid Kruglyak, James Cheverud, Jason Wolf, etc.). GWAS has not failed, it has had remarkable success in the short period of time (or, short at least compared to the period of time where classic Mendelian diseases have been studied) that it has been used in human genetics...and it has also revealed that heritable traits exhibit more complexity than has previously been appreciated. It seems to me that many of the people who search frantically for the "missing heritability" are unfamiliar with most of the literature regarding epistasis in model organisms (since it has hardly been studied at all in humans). For at least some traits (morphological traits and obesity, in particular) there is evidence for a very substantial contribution to heritability from epistatic interactions. In others words, there is a very complex genetic architecture underlying many such traits. We aren't talking about 10 or even 20 genes. We're talking possibly 100s...and just examining the direct additive effects (as is primarily done in human GWAS studies) will not address the proportion of heritability that can be at least partially explained by additive-by-additive, additive-by-dominance and dominance-by-additive types of epistasis. This was a major topic of discussion at the 2010 Biology of Genomes conference at Cold Spring Harbor. The problems that are often touted about GWAS (lack of replication, lack of results) can be at least partially attributable to the following: (1) poor choice of cases (ie: for diseases with too loose a definition/characterization), (2) much too low a sample size (really, for low effect sizes, the power to detect effects is quite low at the typical 2000-5000 sample size level...the real sample sizes for direct effects should be a MINIMUM of ~10,000, and for epistasis, you need at least double that if not 5X that number of samples), (3) population effects. I'm not saying that you cannot detect anything from GWAS with low sample sizes...just look at the cannonical GWAS on AMD by Klein et al. But the cases must be very carefully chosen and the contribution from any single gene detected must be at least moderate for such a study design to work. It's a shame to see articles written and theories espoused as though there can only really be one approach to understanding genetic variation and architecture. I propose that we're likely to find that many Mendelian diseases have some lower level of contribution from other genes ("modifier genes" or small epistatic interaction effects that contribute to severity and penetrance, or age of onset, that kind of thing); complex diseases have some combination of effects from relatively common polymorphisms of sm
More....