Endophagous hymenopteran parasitoids have evolved a variety of strategies to circumvent the immune defences of their insect hosts. An increased number of data on these factors is now available showing that they belong to different types: proteins from wasp venom gland or ovaries, particles of viral origin (polydnaviruses PDVs) and Virus-like particles (VLPs). Some main virulence factors used by parasitoids of dipteran and lepidopteran insects have now been characterized in more or less details (thanks to a previous ANR project an to several collaborations with the Genoscope) as a venom RhoGAP protein from the figitid Leptopilina boulardi, which alters the morphology of immune cells and whose targets in the Drosophila host have been identified. To determine if these factors are common to a majority of species or rather specific and understand the reasons explaining their occurrence in a given parasitoid species, we are using the larger set of data obtained on venom glands (EST from 10 species) and on ovaries (sequences of polydnaviruses genomes produced in the calyx of the ovaries and sequences of EST of ovaries). Preliminary results confirm that the use of one or the other virulence factor can not be deduced from the phylogeny of the wasp or its host taxon and they allow defining less than ten families of factors that can be chosen to perform advanced mechanistic and evolutionary studies. New questions have arised regarding the origin of these eukaryotic virulence factors including the molecular changes leading to transformation of an endogeneous protein into a functional toxin and the way these toxins are delivered into the host (viral symbiosis, secretory system as exosomes ). We then propose here a multidisciplinary approach to examine the diversification of venom toxins across the full range of the 160 million-year-old clade of parasitoid Hymenoptera. Molecules produced in the venom or encoded by PDV all belong to 'classical' eukaryotic protein families. If PDV-encoded factors are directly delivered inside host cells, venom-produced molecules have to be expressed in cells of the venom gland and then secreted into the lumen to be injected along the egg. Most often, cDNAs encoding proteins are also specifically overexpressed in the venom gland compared to the rest of the body (for instance, LbGAP, the Lb serpin LbSPNy, SODs, AGAs, neprilysins are overexpressed in the venom gland). Data are now available on virulence factors introduced in insect hosts by parasitoids but, surprisingly, little is known on the factors produced by hymenoptera gall inducers, except that they might combine maternal secretions initiating gall-formation and egg/larval products ensuring the next steps. Therefore, exploration of the diversity of parasitoid virulence molecules transferred into the insect host, will be extended to a comparison with molecules secreted by gall-forming parasitoids to modulate plant physiology, in an attempt to understand the fundamental bases for evolution of parasite-produced modulating factors for plant and animal.