Decomposed Direct Preference Optimization for Structure-Based Drug Design

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cheng, Xiwei, Zhou, Xiangxin, Yang, Yuwei, Bao, Yu, Gu, Quanquan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910016950239232
author Cheng, Xiwei
Zhou, Xiangxin
Yang, Yuwei
Bao, Yu
Gu, Quanquan
author_facet Cheng, Xiwei
Zhou, Xiangxin
Yang, Yuwei
Bao, Yu
Gu, Quanquan
contents Diffusion models have achieved promising results for Structure-Based Drug Design (SBDD). Nevertheless, high-quality protein subpocket and ligand data are relatively scarce, which hinders the models' generation capabilities. Recently, Direct Preference Optimization (DPO) has emerged as a pivotal tool for aligning generative models with human preferences. In this paper, we propose DecompDPO, a structure-based optimization method aligns diffusion models with pharmaceutical needs using multi-granularity preference pairs. DecompDPO introduces decomposition into the optimization objectives and obtains preference pairs at the molecule or decomposed substructure level based on each objective's decomposability. Additionally, DecompDPO introduces a physics-informed energy term to ensure reasonable molecular conformations in the optimization results. Notably, DecompDPO can be effectively used for two main purposes: (1) fine-tuning pretrained diffusion models for molecule generation across various protein families, and (2) molecular optimization given a specific protein subpocket after generation. Extensive experiments on the CrossDocked2020 benchmark show that DecompDPO significantly improves model performance, achieving up to 95.2% Med. High Affinity and a 36.2% success rate for molecule generation, and 100% Med. High Affinity and a 52.1% success rate for molecular optimization. Code is available at https://github.com/laviaf/DecompDPO.
format Preprint
id arxiv_https___arxiv_org_abs_2407_13981
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Decomposed Direct Preference Optimization for Structure-Based Drug Design
Cheng, Xiwei
Zhou, Xiangxin
Yang, Yuwei
Bao, Yu
Gu, Quanquan
Biomolecules
Machine Learning
Diffusion models have achieved promising results for Structure-Based Drug Design (SBDD). Nevertheless, high-quality protein subpocket and ligand data are relatively scarce, which hinders the models' generation capabilities. Recently, Direct Preference Optimization (DPO) has emerged as a pivotal tool for aligning generative models with human preferences. In this paper, we propose DecompDPO, a structure-based optimization method aligns diffusion models with pharmaceutical needs using multi-granularity preference pairs. DecompDPO introduces decomposition into the optimization objectives and obtains preference pairs at the molecule or decomposed substructure level based on each objective's decomposability. Additionally, DecompDPO introduces a physics-informed energy term to ensure reasonable molecular conformations in the optimization results. Notably, DecompDPO can be effectively used for two main purposes: (1) fine-tuning pretrained diffusion models for molecule generation across various protein families, and (2) molecular optimization given a specific protein subpocket after generation. Extensive experiments on the CrossDocked2020 benchmark show that DecompDPO significantly improves model performance, achieving up to 95.2% Med. High Affinity and a 36.2% success rate for molecule generation, and 100% Med. High Affinity and a 52.1% success rate for molecular optimization. Code is available at https://github.com/laviaf/DecompDPO.
title Decomposed Direct Preference Optimization for Structure-Based Drug Design
topic Biomolecules
Machine Learning
url https://arxiv.org/abs/2407.13981