In recent years, considerable progress has been made in understanding the genomic basis of autism spectrum disorder (ASD). Hundreds of variants have been proposed as predisposing to ASD, and the challenge now is to validate candidates and to understand how gene networks interact to produce ASD phenotypes. Genome-wide association and second-generation sequencing studies in particular have provided important indications about how to understand ASD on a molecular level, and we are beginning to see these experimental approaches translate into novel treatments and diagnostic tests. We review key studies in the field over the past five years and discuss some of the remaining technological and methodological challenges that remain.