Sherlock Hu | CIO
GV20 Therapeutics
Dr. Xihao “Sherlock” Hu is Chief Information Officer and a founding member of GV20 Therapeutics, a biotech company that uses AI to design novel antibody drugs. He holds a Ph.D. in Computer Science from the Chinese University of Hong Kong and completed postdoctoral training at Dana-Farber Cancer Institute and Harvard University. As an interdisciplinary scientist, he develops AI/ML models for target identification and antibody design, work that drove the advancement of GV20-0251 — a first-in-class anti-IGSF8 antibody — into clinical trials. He is the creator of the GV20 STEAD-AI platform for de novo design of fully human functional antibodies, and at GV20 he leads the bioinformatics and AI teams.
Appearances:
Festival of Biologics Day 2 @ 16:50
From AI Target Discovery to Patient Clinical Response in Three Years — The IGSF8 Story
The AI: STEAD / PhialBCR. A permutation-invariant deep-learning model that learns target specificity directly from B-cell receptor (BCR) repertoires computationally assembled from tens of thousands of tumor samples — enabling concurrent target and antibody discovery without protein structure modeling or pre-existing antigen databases.
The discovery: IGSF8 as a novel innate immune checkpoint. PhialBCR prioritized IGSF8 from patient-derived antibody repertoires; orthogonal TCGA genomics and a CRISPR NK-cytotoxicity screen converged on the same target. The same AI then designed GV20-0251, a fully human anti-IGSF8 antibody, from the tumor-derived BCR pool.
The clinical readout: NCT05669430 (n=42). GV20-0251 monotherapy was well tolerated with no DLTs; in anti-PD-1 primary-resistant metastatic melanoma, 3 of 9 evaluable participants achieved confirmed partial responses (ORR 33.3%), with one response ongoing >20 months — AI target prediction to first patient clinical response in ~3 years versus the typical 10–15.
Created by
Terrapinn is proud to be a member of isla.
Working together to build sustainable events