Many biological traits have a significant but complex genetic basis. Identification of causal genetic variants leads to a better understanding regarding the mechanisms of biological trait differences. As a model for complex traits, we are studying an unintentional metazoan experimental evolution experiment performed in C. elegans. Two laboratory domesticated C. elegans strains, N2 and LSJ2, share a common ancestor following its isolation from the wild in 1951. The N2 strain was cultured on agar plates seeded with bacteria for ~11 years while the LSJ2 was cultured in axenic liquid media ~50 years. We have previously found that mutations in
npr-1 and
glb-5 were fixed in the N2 strain, resulting in the suppression of aggregation and bordering behavior to avoid high O2 levels on agar plate. Using a digital PCR-based competition assay, we quantify the relative fitness effects of
npr-1 and
glb-5 ancestry/derived alleles. We find epistatic interaction between these two derived alleles: while the
npr-1 derived allele shows fitness advantages in agar plate conditions regardless of genetic background, the
glb-5 derived allele enhances the
npr-1 fitness effect only in the derived
npr-1 background. Unexpectedly, the fitness benefits of these two alleles seems to occur independently of their effects on behavior as the N2 strain showed increased fitness at 3% and 10% O2 levels where aggregation and bordering is suppressed. To systematically study the potential G x G and G x E in these two strains, we performed fitness QTL mapping using inbred lines (RILs) from N2*(
npr-1 glb-5 ancestry alleles) x LSJ2 on agar plates under two culture conditions: non-starving and starving. We found one QTL with a strong effect on fitness centered over the chromatin remodeling factor
nurf-1. Additionally, we found two QTLs showing a GxE interaction effect. Interestingly, despite lacking the derived alleles of
npr-1 or
glb-5, we found one of the RILs had a high relative fitness equal to the N2 strain. This RIL strain showed changes to bordering and aggregation behaviors distinguishable from the N2 strain. The potential explanations for this high fitness outlier could be due to the fixation of a de novo mutation or higher ordered epistasis of specific combinations of N2 and LSJ2 alleles. We will report the result of our mapping experiments to distinguish between these two possibilities. This study implies experimental evolution approach could be a good method to understand metazoan complex traits and evolution.