Population complexity trumps model complexity in understanding trait variation

The study of expression quantitative trait loci (eQTL) through the use of recombinant inbred lines has yielded detailed information about the transcriptional regulation of complex traits. However, it has proven difficult to apply more advanced genetic models explaining genetic variation underlying gene expression differences. Here, we make use of the difference in genetic complexity of two types of inbred population in the nematode Caenorhabditis elegans to estimate the number of loci affecting gene expression.
We measured gene-expression in a recombinant inbred line (RIL) and an introgression line (IL) population constructed from crossing the strains N2 and CB4856. Both populations received a heat-shock treatment and gene-expression profiles were obtained before (48h at 20oC), directly after heat-shock (2h at 35oC), and after a recovery period (2h at 20oC). Making use of the difference in genetic make-up between the populations - few loci from one parent in the IL versus many in the RILs - allowed for the identification of transcripts regulated by multiple loci. By measuring the transcript variance within each population, for over 1,000 genes across the three conditions we found strong evidence for multiple eQTL underlying gene expression variation. Importantly, most of these multi-loci eQTL are environment-specific. Furthermore, we observed over 200 genes where the phenotypic variation in the IL panel significantly exceeded that in the RIL panel, suggesting evidence for complex genetic buffering.
In conclusion, by using two types of inbred populations the complexity of trait architectures can be investigated without reliance on models of higher complexity. The genetic complexity of a trait is directly observed, rather than estimated post-hoc. Therefore, relying on population complexity rather than model complexity can provide valuable insight in the architecture of quantitative traits.