Webinar: Useful Generalization is the Benchmark That Matters
Wednesday, September 2, 2026 · 1:00 PM PT / 4:00 PM ET
Biotech companies are not trying to win a docking benchmark—they are trying to find chemically novel, experimentally useful starting points against targets that matter. In this webinar, Deep Origin Chief Business Officer Natalie Ma will examine how drug discovery teams and buying committees can become constrained by product categories that do not reflect the scientific problem they are actually trying to solve, and why the most effective systems may need to cross traditional methodological boundaries. Chief Scientific Officer Garegin Papoian will then present the results of Deep Origin’s recent paper, which combines AI-driven pose generation and ranking with physics-based refinement to identify new molecular matter, including on challenging and unfamiliar targets. The session will conclude with a live Q&A on the scientific results, their practical implications, and how teams should evaluate virtual screening technologies around real discovery outcomes rather than isolated benchmarks.
Speakers
Natalie Ma, Ph.D.
Chief Business Officer, Deep Origin
Natalie Ma is Chief Business Officer at Deep Origin, which is building computational discovery systems that model life to close the gap between preclinical predictions and clinical outcomes. She leads business development, strategic partnerships, and government contract programs, including the company's ARPA-H CATALYST award developing next-generation ADMET prediction models. Prior to Deep Origin, Natalie held roles spanning business strategy, partnerships, and commercial operations across the life sciences. She holds a deep interest in the intersection of computational biology, AI, and translational drug development; when not focused on Deep Origin, she spends her time foraging or working on ways to help humans thrive beyond this planet.
Garegin Papoian, Ph.D.
Chief Scientific Officer, Deep Origin
Dr. Garegin Papoian is the Chief Scientific Officer of Deep Origin, with expertise spanning quantum chemistry, protein folding and dynamics, chromatin biophysics, cytoskeletal modeling, and whole-cell simulations. A former Monroe Martin Professor at the University of Maryland and Founder of Biosim (merged with Deep Origin), he earned his PhD at Cornell University under Nobel Laureate Dr. Roald Hoffmann. Dr. Papoian is widely recognized for his pioneering work in computational chemistry and biophysics, with more than one hundred publications in leading international journals and national honors including the Beckman Young Investigator Award, Camille Dreyfus Teacher-Scholar Award, and NSF CAREER Award.
The research behind this webinar
This session presents the results of Overcoming the accuracy-generalization tradeoff in docking and scoring for prospective virtual screening, which introduces DODock and DOScore — ML/physics hybrid frameworks for docking and scoring that generalize out of distribution far better than prior machine-learning approaches.
