A current mystery impacting genome-wide association (GWA) studies is the problem of “missing heritability”, where the total variation explained by all of the candidate loci identified in GWA analyses to date can explain only a small percentage of the heritability of complex phenotypes. There are many possible explanations for why heritability is “missing” (see the following review). One possibility is that there are a large number of genetic loci that contribute to trait heritability, each of which has an effect that is too small to be easily detected with typical GWA studies. We are taking an experimental approach to assess this hypothesis and to develop methodologies for detecting candidate loci with effects small enough that they are “invisible” to a typical GWA study.

The intuitive motivation behind this work is that loci affect a phenotype by altering a pathway or network. Loci may therefore have (relatively) large effects on aspects of critical pathways, while having a small effect on a downstream phenotype. We may therefore be able to discover such loci using a GWA approach, if we analyze the right component of the pathway, i.e. if we analyze a trait affected by the locus, which is important for how the pathway produces variation in the downstream phenotype.

The model system we are using for this work is body size in Drosophila melanogaster, which we quantify by total body weight and by wing size. While variation in body size is complex, there are characterized pathways where experimental alteration of the pathway has been demonstrated to alter body size. We are using such pathways as a guide to determine which traits to measure. The entire project involves a GWA analysis of both body size and relevant pathways using lines of the Drosophila Genetic Reference Panel (DGRP), and using network modeling tools to establish the connections between pathway components and size. This work therefore makes use of our computational genome-wide association, network discovery, and medical genomics techniques and provides an integrative foundation for our research.





Output of the automated system that we use to measure wing size (and cell number), developed by Dr. David Houle at Florida State University.
Results quantifying the relationship between expression of one of the insulin-like proteins (ILP*) and body / wing size in the DGRP lines measured using rtPCR (top). Results of a GWA on the gene expression level of ILP* using SNPs called by our group for DGRP (bottom).