The genome of B. jararaca was sequenced and is publically available. It allowed the identification of canonical gene structures of 14 families of toxins, the most complete set of snake toxin genes described so far. Besides their intron-exon arrangement, we assigned the genomic locus for most of these genes, which showed that some toxin families (e.g. SVMP and SVSP) have paralogous genes occurring in tandem, whereas other exist as single copy genes (e.g. BPP and VEGF- A). Moreover, the genomic context provided important information about the origin of some toxins. SVMP-genes are flanking the ADAM28 gene (their supposed ancestor), indicating that they probably originated from the duplication of this gene. The VEGF-F-gene occupies the position of a PGF-like in other animals, indicating that, contrarily to previous suppositions, this toxin has not arisen by duplication and neofunctionalization of the VEGF-A gene and more likely resulted from a direct gene recruitment.