Language AI for RNA Virus and RNA Vaccine
Liang Huang (Oregon State University | Coderna.ai)
Distinguished Lecture
Thursday, November 30, 2023, 3:30 pm
Abstract
Linguistics and biology are two sides of the same coin. This talk features several highly unexpected connections between them which yield efficient algorithms with substantial biological impacts. One such connection (Nature, 2023) is between messenger RNA (mRNA) vaccines and formal language theory. Although widely used in COVID, these vaccines still suffer from instability. But how to design more stable and efficient mRNAs? Here we show a surprising reduction of the mRNA design problem to the classical (1961) concept of “lattice parsing” in speech recognition, which enables efficient search in the exponentially large design space. Experiments on COVID and another virus show that our designs dramatically improves mRNA half-life, protein expression, and in vivo antibody response, compared to the standard method used by Pfizer and Moderna. Another connection (PNAS, 2021) is between COVID variants and multilingual parsing. Here we show that aligning and folding various coronavirus genomes (in order to find conserved structures for drug design) can be viewed as “synchronous parsing” for multiple languages. This enables efficient global prediction of COVID genome structure that matches experimental work.
[1] Nature paper: https://www.nature.com/articles/s41586-023-06127-z
[2] Nature news: https://www.nature.com/articles/d41586-023-01487-y
(‘Remarkable’ AI tool designs mRNA vaccines that are more potent and stable)
[3] PNAS paper: https://www.pnas.org/doi/10.1073/pnas.2116269118
Bio
Liang Huang (Ph.D., Penn, 2008) is a Professor of Computer Science at Oregon State University, and co-founder of Coderna.ai. Until recently, he was also a Distinguished Scientist at Baidu Research USA. He also worked at Google Research, USC, and City Univ. of New York. He was known for algorithms and theory in computational linguistics, where he received several best paper awards (ACL 2008 Best Paper Award, EMNLP 2016 Best Paper Honorable Mentions, NAACL 2022 Best Demo Paper Award) and delivered keynotes at ACL 2019 and CVPR 2021. But in recent years, he has shifted his attention to applying these natural language algorithms to computational biology, esp. RNA folding and RNA design, with the hope of fighting COVID. This line of linguistics-inspired biology work eventually led to PNAS (2021) and Nature (2023) papers, and is widely covered in the media.
This talk is available on the Allen School's YouTube channel.