For decades following the modern synthesis in evolutionary biology, the molecular details underlying the genotype-phenotype map have been treated as a "black box." Propelled by new technologies, our ability to peer inside the box is rapidly improving. Using a systems genetics framework, we can now integrate genotypic and phenotypic variation with variation in molecular endophenotypes (such as gene expression) to construct biological networks that allow us to draw the causal path from variation in allele frequencies to phenotypic differences between individuals. However, the average allelic effect across individuals does not always capture the effect of a specific allele under certain conditions. For example, both the genetic background and environmental differences can lead to major phenotypic differences among individuals sharing the same alleles. When this occurs, current methods may not be able to detect the link between genotype and phenotype -- this may help explain why we have fallen short of the ultimate goal of explaining most of the heritability for complex traits in terms of allelic effects. My research is dedicated to understanding the context dependence of allelic effects in light of gene-gene and genotype-by-environment interactions. While we have historically thought of the genotype-phenotype map as a static blueprint underlying the phenotype, the work I will present suggests that the blueprint itself is dynamic and context-dependent. This perspective is particularly helpful when trying to understand, for example, why some individuals appear to be more sensitive than others to environmental perturbations. I will illustrate this point using examples from my current work, and I will discuss the implication of these findings in the context of evolutionary response and our understanding of the genetic basis of complex traits.