高水平男子中短距离游泳成绩预测的代谢组学模型

Metabonomical models for predicting the performances of high performance male medium and long distance swimmers

为了探索利用代谢组学方法预测高水平男子中短距离游泳成绩的可能性。第15届亚运会赛前1个月,每周1次,利用基于核磁共振(NMR)的代谢组学方法对18名准备参加亚运会的男子运动员晨尿进行分析,然后采用有监督的模式识别方法——Fisher逐步判别分析法,根据从一维核磁共振氢谱获取的标准化数据,建立高水平男子中短距离游泳成绩预测代谢组学模型。结果显示:模型的预测性能检验结果,对比赛中奖牌选手(进入前3名)与非奖牌选手(未进入前3名)、决赛选手(进入前8名)与非决赛选手(未进入前8名)的样本预测准确率可达到90%以上。初步证明了利用代谢组学模型预测高水平男子中短距离游泳比赛成绩的可行性。

In order to probe into the possibility to predict the performances of high performance male medium and long distance swimmers by utilizing metabonomical methods,the authors analyzed the morning urine of 18 male swimmers preparing for participating in the Asian Games once a week,1 month before the 15th Asian Games,by utilizing metabonomical methods based on nuclear magnetic resonance（NMR）,then established metabonomical models for predicting the performances of high performance male medium and long distance swimmer based on standardized data obtained from the one dimensional NMR spectrum of hydrogen by using a monitored mode iden-tification method – the gradual discriminant analysis method created by Fisher,and revealed the following test result of the prediction performance of the models： the sample prediction accuracy rates of the medalists（top 3）,non medalists（beyond top 3）,finalists（top 8） and non finalists（beyond top 8） were up to 90%,which preliminarily proved the feasibility to predict the competition performances of high performance male medium and long distance swimmers by utilizing metabonomical methods.