基于双层生物分子网络的致病基因预测

Disease Gene Prediction Based in Two-layer Biomolecular Network

基于“致病基因邻居倾向于引起相似疾病”和“必需基因和致病基因关联较少”的假设,融合蛋白质相互作用和代谢通路,构建双层生物分子网络,提出一种基于双层生物分子网络双流量网络传播的致病基因排序方法。为致病基因和必需基因分配正流量和负流量,挖掘与已知致病基因相似的潜在致病基因。实验结果表明,本方法与基于经典假设的单层网络方法相比,在富集分数1和富集分数2分别提高11.66和13.66。

Based on the hypothesis that"the neighbors of pathogenic genes tend to cause similar diseases"and"the essential genes and disease genes are less related",a two-layer biomolecular network was constructed by fusing protein interaction and metabolic pathways.A method of sequencing disease genes was proposed based on two-layer biomolecular network and dual flows network transmission.Positive flow and negative flow were allocated for disease genes and essential genes and potential pathogenic genes similar to known pathogenic genes were mined.Experimental results show that this method compared with single network method based on the classical assumptions,the enrichment score 1 and enrichment score 2 increased by 11.66 and 13.66 respectively.