融合原子交换特征信息的代谢路径预测
Metabolic pathway prediction using feature information of atom exchange
黄毅然 1万志远 2钟诚1
作者信息
- 1. 广西大学计算机与电子信息学院,广西南宁 530004;广西大学广西高校并行与分布式计算技术重点实验室,广西 南宁 530004;广西大学 广西多媒体通信与网络技术重点实验室,广西 南宁 530004
- 2. 广西大学计算机与电子信息学院,广西南宁 530004
- 折叠
摘要
为获取从任意起始代谢物到目标代谢物的生化相关性较好的代谢路径,提出一种融合原子交换特征信息的基于约束的代谢路径预测算法PVA.结合代谢网络中具有的原子交换特征信息,建立一种基于约束的混合整数线性规划(MILP)代谢路径预测模型,以搜索从任意起始代谢物到目标代谢物并包含特定原子交换信息的代谢路径.实验结果表明,与同类方法相比,PVA能够有效地发现生化相关性更好的代谢路径.
Abstract
To search the metabolic pathways with better biochemical correlation from an arbitrary starting metabolite to the given target metabolite,the constraint-based method called PVA was proposed to predict metabolic pathways by fusing atom exchange feature information.The atom exchange feature information in the metabolic network was combined to construct a constraint-based mixed integer linear program(MILP)model for searching the metabolic pathways containing specific atom exchange from an arbitrary starting metabolite to the given target metabolite.Experimental results demonstrate that compared with existing methods,the proposed method PVA can effectively find better biochemically feasible metabolic pathways.
关键词
代谢网络/代谢路径预测/原子交换/混合整数线性规划/化学计量/路径优化/代谢工程Key words
metabolic network/metabolic pathway prediction/atom exchange/mixed integer linear program/stoichiometry/pathway optimization/metabolic engineering引用本文复制引用
基金项目
国家自然科学基金项目(61862006)
国家自然科学基金项目(61861004)
广西自然科学基金项目(2020GXNSFAA159074)
出版年
2024
NETL
NSTL
万方数据