A different perspective on genetic architecture

As an ecological geneticist, I’m constantly reminded how much we don’t understand about the genetic nature of adaptive variation. Sure, we have lots of examples of genes/pathways/regions that seem to be responsible for adaptation, but we don’t really know if these are representative of typical evolutionary processes. Are large or small effect mutations important? Does selection act on gene regulation or protein coding regions? Moreover, our expectations have changed through time with the development of theoretical models and improved empirical data collection. In his recent manuscript in Evolution, Remington offers a solution to this problem by putting forth a model of allelic evolution that takes a molecular perspective. Before digging into Remington’s idea, a little background is necessary.

The infinitesimal model argues that quantitative traits are controlled by an infinite number of loci of small effect. This model was extended to Fisher’s geometric model, which provides an evolutionary justification for why the infinitesimal model works. Basically, the geometric model argues that only alleles of small effect have any real chance of being beneficial.

However, Remington points out that QTL mapping regularly identifies alleles of large effect, though much variance is still left unexplained and suggests a polygenic (but not infinitesimal) basis of quantitative traits. These findings suggest that alleles of large effect are indeed important and have led us down the path of the search for causative SNPs. But closer examination of many QTL regions show that multiple small effect loci actually underlie the large effect QTL. Moreover, genome wide association studies (GWAS) typically only account for very small proportions of variance using a SNP by SNP approach.

These discrepancies have lead to some scathing critiques of the QTL perspective, arguing that alleles of large effect are not representative of typical evolutionary processes (a worthwhile read on this topic is Rockman 2012). Remington disagrees, and thinks that a more appropriate model of allelic variation is needed.

Remington makes two main arguments:

  1. Allelic effects are important for driving phenotype.
  2. Selection regimes differ within versus between populations and can result in different trait architecture.

By allelic effects, Remington basically means the haplotype. He argues that the typical geometric/infinitesimal perspective relies on the sequential accumulation of mutations during an adaptive walk (see panel A of figure below). In reality, there can be multiple haplotypes harboring different mutations within a population. These mutations may be of small effect when on different haplotypes, but when brought together through recombination may jointly have a large effect on phenotype (panel B of figure), thus producing large effect alleles.


A. Conventional model of adaptive walk. B. Adaptive walk at an allelic level. From Remington 2015

Secondly, he argues that it is important to distinguish between intra-population traits and inter-populations traits. For example, in panmictic populations,

“Adaptively important traits…are likely to be under some form of direct or indirect stabilizing selection, which would tend to reduce the frequency of large-effect alleles since individuals carrying them would be more likely to deviate from optimal phenotypes.”

In contrast, large effect QTLs have generally been found in populations or species under strong divergent selection. Under this scenario, it is reasonable that selection would favor alleles of large effect during adaptation, especially when gene flow is present (Yeaman and Whitlock 2011).

The summary here obviously only scratches the surface of Remington’s broad argument. In general, I think the manuscript offers an interesting perspective on the genetic architecture of quantitative traits. Though, I’d be quite interested to hear any issues with his arguments.


Remington, D.L., 2015. Alleles versus mutations: Understanding the evolution of genetic architecture requires a molecular perspective on allelic origins. Evolution, 69(12), pp.3025-3038.

Rockman, M.V., 2012. The QTN program and the alleles that matter for evolution: all that’s gold does not glitter. Evolution, 66(1), pp.1-17.

Yeaman, S. and Whitlock, M.C., 2011. The genetic architecture of adaptation under migration–selection balance. Evolution, 65(7), pp.1897-1911.


About Reid Brennan

I am an evolutionary ecologist and a postdoc at the University of Vermont. I am generally interested in mechanisms allowing populations to evolve and respond to environmental stressors.
This entry was posted in association genetics, evolution, genomics, mutation, quantitative genetics, theory. Bookmark the permalink.