Enhancing Safe and Controllable Protein Generation via Knowledge Preference Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yuhao, Ding, Keyan, Feng, Kehua, Wang, Zeyuan, Qin, Ming, Li, Xiaotong, Zhang, Qiang, Chen, Huajun
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913941180907520
author Wang, Yuhao
Ding, Keyan
Feng, Kehua
Wang, Zeyuan
Qin, Ming
Li, Xiaotong
Zhang, Qiang
Chen, Huajun
author_facet Wang, Yuhao
Ding, Keyan
Feng, Kehua
Wang, Zeyuan
Qin, Ming
Li, Xiaotong
Zhang, Qiang
Chen, Huajun
contents Protein language models have emerged as powerful tools for sequence generation, offering substantial advantages in functional optimization and denovo design. However, these models also present significant risks of generating harmful protein sequences, such as those that enhance viral transmissibility or evade immune responses. These concerns underscore critical biosafety and ethical challenges. To address these issues, we propose a Knowledge-guided Preference Optimization (KPO) framework that integrates prior knowledge via a Protein Safety Knowledge Graph. This framework utilizes an efficient graph pruning strategy to identify preferred sequences and employs reinforcement learning to minimize the risk of generating harmful proteins. Experimental results demonstrate that KPO effectively reduces the likelihood of producing hazardous sequences while maintaining high functionality, offering a robust safety assurance framework for applying generative models in biotechnology.
format Preprint
id arxiv_https___arxiv_org_abs_2507_10923
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Safe and Controllable Protein Generation via Knowledge Preference Optimization
Wang, Yuhao
Ding, Keyan
Feng, Kehua
Wang, Zeyuan
Qin, Ming
Li, Xiaotong
Zhang, Qiang
Chen, Huajun
Artificial Intelligence
Protein language models have emerged as powerful tools for sequence generation, offering substantial advantages in functional optimization and denovo design. However, these models also present significant risks of generating harmful protein sequences, such as those that enhance viral transmissibility or evade immune responses. These concerns underscore critical biosafety and ethical challenges. To address these issues, we propose a Knowledge-guided Preference Optimization (KPO) framework that integrates prior knowledge via a Protein Safety Knowledge Graph. This framework utilizes an efficient graph pruning strategy to identify preferred sequences and employs reinforcement learning to minimize the risk of generating harmful proteins. Experimental results demonstrate that KPO effectively reduces the likelihood of producing hazardous sequences while maintaining high functionality, offering a robust safety assurance framework for applying generative models in biotechnology.
title Enhancing Safe and Controllable Protein Generation via Knowledge Preference Optimization
topic Artificial Intelligence
url https://arxiv.org/abs/2507.10923