TwoFold: Highly accurate structure and affinity prediction for protein-ligand complexes from sequences

Darren J. Hsu; Hao Lu; Aditya Kashi; Michael Matheson; John Gounley; Feiyi Wang; Wayne Joubert; Jens Glaser

doi:10.1177/10943420231201151

TwoFold: Highly accurate structure and affinity prediction for protein-ligand complexes from sequences

Darren J. Hsu, Hao Lu, Aditya Kashi, Michael Matheson, John Gounley, Feiyi Wang, Wayne Joubert, Jens Glaser

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

We describe our development of ab initio protein-ligand binding pose prediction models based on transformers and binding affinity prediction models based on the neural tangent kernel (NTK). Folding both protein and ligand, the TwoFold models achieve efficient and quality predictions matching state-of-the-art implementations while additionally reconstructing protein structures. Solving NTK models points to a new use case for highly optimized linear solver benchmarking codes on HPC.

Original language	English
Pages (from-to)	666-682
Number of pages	17
Journal	International Journal of High Performance Computing Applications
Volume	37
Issue number	6
DOIs	https://doi.org/10.1177/10943420231201151
State	Published - Nov 2023

Funding

This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-00OR22725.

Funders	Funder number
Office of Science	DE-AC05-00OR22725

Keywords

COVID-19
drug design
machine learning
neural tangent kernel
transformer

Access to Document

10.1177/10943420231201151

Cite this

@article{f6fb863d1b1b4daababc6d07b7d964d4,

title = "TwoFold: Highly accurate structure and affinity prediction for protein-ligand complexes from sequences",

abstract = "We describe our development of ab initio protein-ligand binding pose prediction models based on transformers and binding affinity prediction models based on the neural tangent kernel (NTK). Folding both protein and ligand, the TwoFold models achieve efficient and quality predictions matching state-of-the-art implementations while additionally reconstructing protein structures. Solving NTK models points to a new use case for highly optimized linear solver benchmarking codes on HPC.",

keywords = "COVID-19, drug design, machine learning, neural tangent kernel, transformer",

author = "Hsu, {Darren J.} and Hao Lu and Aditya Kashi and Michael Matheson and John Gounley and Feiyi Wang and Wayne Joubert and Jens Glaser",

note = "Publisher Copyright: {\textcopyright} The Author(s) 2023.",

year = "2023",

month = nov,

doi = "10.1177/10943420231201151",

language = "English",

volume = "37",

pages = "666--682",

journal = "International Journal of High Performance Computing Applications",

issn = "1094-3420",

publisher = "SAGE Publications",

number = "6",

}

TY - JOUR

T1 - TwoFold

T2 - Highly accurate structure and affinity prediction for protein-ligand complexes from sequences

AU - Hsu, Darren J.

AU - Lu, Hao

AU - Kashi, Aditya

AU - Matheson, Michael

AU - Gounley, John

AU - Wang, Feiyi

AU - Joubert, Wayne

AU - Glaser, Jens

N1 - Publisher Copyright: © The Author(s) 2023.

PY - 2023/11

Y1 - 2023/11

N2 - We describe our development of ab initio protein-ligand binding pose prediction models based on transformers and binding affinity prediction models based on the neural tangent kernel (NTK). Folding both protein and ligand, the TwoFold models achieve efficient and quality predictions matching state-of-the-art implementations while additionally reconstructing protein structures. Solving NTK models points to a new use case for highly optimized linear solver benchmarking codes on HPC.

AB - We describe our development of ab initio protein-ligand binding pose prediction models based on transformers and binding affinity prediction models based on the neural tangent kernel (NTK). Folding both protein and ligand, the TwoFold models achieve efficient and quality predictions matching state-of-the-art implementations while additionally reconstructing protein structures. Solving NTK models points to a new use case for highly optimized linear solver benchmarking codes on HPC.

KW - COVID-19

KW - drug design

KW - machine learning

KW - neural tangent kernel

KW - transformer

UR - http://www.scopus.com/inward/record.url?scp=85175292692&partnerID=8YFLogxK

U2 - 10.1177/10943420231201151

DO - 10.1177/10943420231201151

M3 - Article

AN - SCOPUS:85175292692

SN - 1094-3420

VL - 37

SP - 666

EP - 682

JO - International Journal of High Performance Computing Applications

JF - International Journal of High Performance Computing Applications

IS - 6

ER -

TwoFold: Highly accurate structure and affinity prediction for protein-ligand complexes from sequences

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this