Table 1 Possible mechanisms underlying the five largest factors affecting Embedded Image
EffectCategoryMagnitudeMechanistic interpretation
Embedded ImageSpecies (main effect)−1.44The C. muridarum amino acid background is thermodynamically unstable, more prone to proteolytic degradation, and has low catalytic efficiency. Consequently, it has a strong negative effect on the ability to survive in the presence of drug, across all other interacting genetic backgrounds.
L28RSNP (main effect)+1.22The L28R mutation greatly increases both structural stability and the drug inhibition constant Embedded Image, and, consequently, helps DHFR perform its enzymatic function in the presence of drug, across genotypic contexts.
Embedded ImageSpecies (main effect)−0.90The L. grayi amino acid background is very thermodynamically unstable and prone to proteolytic degradation. This is partially compensated for by reasonably high catalytic efficiency Embedded Image, but still has a net negative effect on Embedded Image.
Embedded Image: P21LSpecies × SNP (second-order)−0.82The C. muridarum amino acid background is inefficient and thermodynamically unstable. However, the P21L mutation is slightly stabilizing, which diminishes the negative impact of the C. muridarum amino acid background. The net effect remains negative, however. This result highlights how powerful the C. muridarum amino acid background is, in that it can “drag down” the positive effects of certain SNPs.
Embedded Image:L28RSpecies × SNP (second-order)−0.69This highlights the nonlinear interaction between a powerfully positive SNP (L28R) and the strongly negative main effect L. grayi background. That the interaction term is negative highlights that even the stabilizing effects of a positive effect SNP (L28R) cannot compensate for the negative effects of the unstable L. grayi amino acid background.
  • DHFR, dihydrofolate reductase.