Meeting Abstract
Animal venoms share some striking similarities which have been explained on the light of three concurrent paradigms: 1) the “proteome to venome” model of venom protein emergence, highlighting the close relationship of some known and assumed toxins with “normal” (i.e. non-toxin) proteins from the same or a different organism; 2) the “convergent recruitment and expansion” of a few protein folds into phylogenetically unrelated venoms, in other words, independently evolved venoms are built upon lineage specific expansions of homologous genes, and; 3) a prey-driven “evolutionary arms-race” scenario of venom-protein evolution. These paradigms were tremendously helpful when the high-throughput screenings became routine analysis; for a while, it seemed that almost every new venomics study supported these models. However, this proliferation of venom surveys, with no or little reference to venom ecology and organismal biology, has the intrinsic risk of propagating annotation errors and a tendency of ecological over-interpretation based on undersampled phylogenies. Here, based on de-novo assembleges of over twenty deep-sequenced scorpion transcriptomes and surveys for toxin/toxin-like genes, we will show that: 1) few protein families have evolutionary histories which are congruent with the current paradigms of venom evolution; 2) most of the diversity and selection on these protein families is restricted to a single scorpion lineage, and; 3) a single scorpion family has evolved an over-kill venom, likely driven by few major ecological shifts. We call for a revision of the current (scorpion) venom evolution paradigms to a model where strong directional selection pressure results in sparse structural and functional recruitments, while allowing the accumulation of paralogs under neutral selection.
