The science behind Cavatar's caveolae-targeting platform

When we say AVATAR's models are "trained on 20+ years of in-vivo-validated data," it's worth being precise about where that data actually came from — because in a field full of AI platforms trained on public databases and inferred correlations, the provenance of training data is the difference between a model that's confident and a model that's right.

That data has two distinct origins. The earliest foundational work — establishing that caveolae actively transport cargo across the endothelium at all — was Jan Schnitzer's own academic research in the 1990s and 2000s (Schnitzer et al., J Cell Biol, 1994; McIntosh et al., PNAS, 2002; Oh et al., Nature, 2004; Oh et al., Nature Biotechnology, 2007). In 2009, Schnitzer founded PRISM (Proteogenomics Research Institute for Systems Medicine), a nonprofit research institute, to continue and scale that work — mapping the vascular surface tissue by tissue, and validating specific delivery routes in the lung (Chrastina et al., Journal of Vascular Research, 2010) and in solid tumors (Oh et al., Nature Medicine, 2014), later extending to an anti-fibrotic therapeutic at ultra-low doses (Kadam et al., PLOS ONE, 2022) and to nanoparticle cargo (Nayak, Chrastina et al., Nature Nanotechnology, 2024).

Several of the researchers behind that body of work — including Adrian Chrastina, Michael Levin, and Bogdan Olenyuk — are part of Cavatar's team today. Founded in 2016, Cavatar is an independent company built to take that scientific lineage and turn it into something PRISM's nonprofit research mission was never designed to do: a commercial, AI-driven platform that can rank targets for any biotech, at scale.

What this means when AVATAR ranks a target

Every score AVATAR produces — druggability, tissue accessibility, delivery-format fit — is anchored to that lineage, not inferred from expression databases alone. When the platform calls a target "accessible," that claim is backed by the same category of peer-reviewed, in-vivo evidence that took Schnitzer's original discovery from a 1994 mechanism to a validated, tissue-specific delivery route — carried forward by the same people who did that work, now building the tools to do it faster.

Read the research behind AVATAR: Schnitzer et al. (1994) · McIntosh et al. (2002) · Oh et al. (2004) · Oh et al. (2007) · Chrastina et al. (2010) · Oh et al. (2014) · Kadam et al. (2022) · Nayak, Chrastina et al. (2024)