一种新的用于候选基因排序的数据融合方法

Novel data fusion method for candidate gene prioritization

从成百上千的候选基因中确定关键基因是寻找致病基因(或参与某个生物过程的基因)的重要步骤,而根据多种数据源对候选基因进行综合排序则成为该领域新的挑战。提出一种新的基于单类支持向量机的数据融合方法用于候选基因排序。实验表明该方法可以有效地利用多种异构的生物数据源对候选基因排序,其准确率和鲁棒性均优于根据单数据源进行排序。

Identifying key candidates in the thousands of genes in a genome is an important step in hunting genes playing roles in a disease phenotype or a complex biological process, and candidate gene prioritization integrating kinds of data sources is becoming a new challenge in this field. A new data fusion method based on one-class Support Vector Machine （SVM） was proposed for candidate gene prioritization. Experimental results indicate that the proposed method is valid in gene prioritization integrating kinds of heterogeneous data sources and its accuracy and robustness are better than that of the method with single data source.