【摘要】 将一组Hela细胞的时序数据通过傅立叶分析变换为傅立叶谱,并设计了基于支持向量机的有监督学习算法,该算法通过提取傅立叶谱中的显著特征鉴定持家基因和非持家基因。本文所提出的方法通过比较两套独立的组织表达谱成功预测了510个人类持家基因,其中包括93个非编码持家基因。分析结果表明:本文方法所预测的持家基因相比其他3种方法更为高效、准确。更多还原

【Abstract】 A group of time series data of Hela cells is transformed to Fourier spectrum by Fourier analysis.A support vector machine based monitoring learning algorithm is proposed.This algorithm is applied to pick out the Housekeeping IncRNAs from the Fourier spectrum,which can extract important features of the system,providing a basis for identifying the specific RNAs expression patterns.Using the above method 510 human Housekeeping genes are confirmed,which are then identified by comparison with two standard sets of human tissue specific expression profiles.Results show that the proposed method can give more reliable Housekeeping IncRNAs than three existing identifying methods.更多还原