摘要
本文运用神经网络方法，并结合分子生物学的有关理论与实验统计事实，对真核基因启动子区域进行了识别．文中选择了人类、牛、猪、猫、山羊、兔、绵羊、大鼠、小鼠、小马、仓鼠、鸡、鸭、大豆等共13种真核生物360个基因组，作为研究对象．学习组选择了300个基因组，预测组选择了60基因组．结果表明，将神经网络模型与基因理论相结合，能够运用计算方法，从大量的可能启动子组合中排列出唯一的启动子区域．
In this paper,a neural network method(including molecubiology theory and expermental statistical facts)was applied to predict the promoter locations in eukaryotic gene.360 genes of 14 kinds of organisms including human,bull,pig,cat,goat,rabbit,sheep,rat,mouse,house,hamster,chicken,duck and bean were selected as an object of study.In thses genes,300 genes were selected as the training set to construct a net-work model and other 60 genes were used as the samples for prediction.As a result,by using neural network method and gene thegry,the position of promoter could be obtained by computer from numerous possible combinations of predicted promoters.
出处
《生物数学学报》
CSCD
北大核心
1994年第1期77-80,共4页
Journal of Biomathematics
基金
中科院院长基金会资助
关键词
真核基因
启动子
人工神经网络
Eukaryote Gene,Promoter,TATA box.CAAT box,GC box,Artificial Neural Network,Back-Propagation Model.
