© Matías Estarás Hermosel

Hugo Bisio SabarisScientist

ERC Proof of Concept

Information génomique et structurale (IGS)1

PSDisrupt (Horizon Europe - ERC PoC 101247194)
AI-driven discovery broad-spectrum antivirals targeting Nucleocytoviricota factories via phase separation modulation

The phylum Nucleocytoviricota includes (re-)emerging dsDNA viruses of major concern, such as monkeypox virus (mpox) and African swine fever virus (ASFV). Despite their economic and public health impact, no formally approved antivirals exist for these pathogens.

Viral factories (membrane-less organelles essential for replication) are highly conserved across Nucleocytoviricota, presenting an attractive and untapped target for antiviral intervention.

PSDisrupt builds on findings from the ERC-funded ViDaMa project, which identified scaffold proteins governing phase separation in viral factories. These proteins share a conserved "molecular grammar" essential for condensate formation. Disrupting this process is expected to broadly inhibit viral replication while minimizing the development of resistance.

This PoC project will design phase-separation grammar-targeting antivirals using AI-driven pipelines. Deep learning models will enable the screening of ultra-large chemical spaces to identify: (i) broad-spectrum inhibitors targeting conserved scaffold proteins across Nucleocytoviricota, and/or (ii) virus-specific antivirals for mpox and ASFV.

PSDisrupt’s translational goal is to design and advance lead compounds into preclinical validation and assess their commercial potential. We will:

  • Discover and design drugs by leveraging high throughput screens and deep learning-based algorithms.
  • Perform biophysical and cellular assays to validate compound efficacy.
  • Establish IP protection for novel inhibitors.
  • Engage with pharmaceutical and biotech partners for licensing and further development.

By bridging virology and AI-driven drug design, PSDisrupt pioneers a novel antiviral strategy with broad applications beyond infectious diseases. The AI framework could be adapted to target phase separation in diverse pathological conditions, including cancer, neurodegeneration, and other condensate-associated diseases, significantly expanding its biomedical and commercial impact.

  • 1Aix-Marseille Université/CNRS
-A +A
Contact
Print