|国家科技期刊平台
首页|期刊导航|中山大学学报(自然科学版)(中英文)|基于嗅觉受体激活关系模拟的气味感知预测

基于嗅觉受体激活关系模拟的气味感知预测OACSTPCD

Prediction of olfactory perception based on simulation of olfactory receptor activation relationships

中文摘要英文摘要

气味分子与嗅觉受体相互作用是引起气味感知的重要环节,对于揭示气味感知机制具有重要意义.然而,获得气味分子与人类嗅觉受体激活关系的实验性结果耗时耗力,且目前可用的激活关系数据数量不足以支持智能气味感知研究.因此,本研究构建了嗅觉受体蛋白质关系网络,并提取特征来训练气味分子-嗅觉受体激活关系预测模型.在气味感知预测中综合考虑气味分子特征和嗅觉受体蛋白激活模拟关系,实现了对人类气味感知的高精度回归预测.实验结果表明,融合气味分子-嗅觉受体激活关系的人类气味感知预测相关度指标为0.94,明显优于现有的气味感知预测模型.此外,研究还在预测基础上总结了气味分子-嗅觉受体激活-气味感知模式.本研究为气味感知预测引入了可观测的嗅觉受体激活机制特征,为深入探索和理解气味感知机制提供了新思路.

The interaction between odor molecules and olfactory receptors is a crucial step in olfactory perception and holds significant importance in unraveling the mechanism of olfactory perception.How-ever,obtaining experimental results on the activation relationship between odor molecules and human olfactory receptors is time-consuming and labor-intensive,and the available data on activation relation-ships is currently insufficient to support intelligent olfactory perception research.Therefore,this study constructed a network of olfactory receptor protein relationships and extracted features to train a model for predicting the activation relationship between odor molecules and olfactory receptors.By integrating the features of odor molecules and the simulated activation relationship of olfactory receptor proteins in olfactory perception prediction,high-precision regression prediction of human olfactory perception was achieved.Experimental results showed that the correlation coefficient of human olfactory percep-tion prediction fused with odor molecule-olfactory receptor activation relationship reached 0.94,signifi-cantly outperforming existing olfactory perception prediction models.Additionally,the study summa-rized the odor molecule-olfactory receptor activation-olfactory perception pattern,enriching our under-standing of the mechanism of smell perception.This study introduced observable features of olfactory receptor activation mechanisms into olfactory perception prediction,providing new insights for further exploration and understanding of the mechanism of olfactory perception.

左敏;胡静珺;颜文婧;王瑞东;张青川;范大维

北京工商大学农产品质量安全追溯技术及应用国家工程研究中心,北京 100048北京市房山区教师进修学校,北京 102401

生物学

分子特征提取蛋白质特征提取嗅觉受体激活预测气味感知预测图卷积机器学习

molecular feature extractionprotein feature extractionolfactory receptor activation predictionolfactory perception predictiongraph convolutionmachine learning

《中山大学学报(自然科学版)(中英文)》 2024 (001)

86-95 / 10

国家重点研发计划项目(2021YFD2100605);北京市属高校教师队伍建设支持计划高水平科研创新团队项目(BPHR20220104)

10.13471/j.cnki.acta.snus.2023E040

评论