Birth defects affect approximately 3% of U.S. infants and are a major cause of morbidity and mortality. Despite their prevalence, little is known about their causes. To discover genomic variants associated with rare, diagnostically challenging birth defects in newborn infants and children, we use next generation sequencing (exome or whole genome sequencing) of affected infant/parent trios. We computationally identify rare variants (minor allele frequency less than 0.01) and predict function with a suite of algorithms that incorporate evolutionary conservation and known regulatory motifs. To prove causality of discovered variants, we use model systems (worms, fruit flies, zebrafish, mice) to demonstrate function of discovered variants, and replication of candidate gene loci or pathways in unrelated infants with similar phenotypes.