Genetics has revolutionized the conceptualization of autism, and promises rationale, mechanistically based therapeutics in the future. At the same time, as our genetic understanding of ASD etiology advances, we must first tackle extraordinary complexity at multiple biological levels, from molecules to cells to circuits and behavior. Recent work in the genetics of ASD predicts that about 1000 genes contribute to the overall population risk for ASD. Furthermore, no major effect gene accounts for more than 1% of cases. Genetic effect sizes range from very small (for common variants), to rare mutations that are essentially considered causal. However, even for the most major, causal mutations, the effects appear to extend across current disease boundaries. These data have challenged our lab to focus on two major questions: understanding what are the neurobiological consequences of ASD risk mutations and whether there are common or convergent molecular mechanisms that link ASD cases with diverse etiologies. I will discuss our work using genetically based animal and in vitro models to understand ASD pathophysiology, as well as our work using systems biology approaches and network inference to connect multiple levels of biology in a coherent manner. Our work in transcriptional profiling reveals strong evidence of convergence on common pathways in the ASD post mortem brain. Overall, these data and that of others support the existence of convergent molecular pathways in ASD. Using this information to develop novel therapeutics in children and adults with ASD remains a critical challenge.