通过文本挖掘获取疾病相关功能信息

Retrieving Gene Functional Information through Text Mining

通过定位候选策略和全基因组关联研究等方法,很多人类遗传疾病的致病基因已经定位到某个或某些染色体区间,利用计算机将染色体区间中众多的基因减少到易于实验分析的数目是寻找疾病基因的一个很重要的方法。大部分已有的预测疾病基因的方法都是利用已知致病基因的各类注释信息来预测疾病基因的。但是,目前依然有很多疾病尚没有任何具体的注释信息,这样就无法利用已有的基于已知基因信息的预测方法来识别致病基因。针对这个问题,通过挖掘生物医学文献数据库,结合人类基因产物蛋白质的功能注释数据库,从中提取与疾病相关的功能信息。这样,就可以基于这些挖掘出来的功能信息来实现这类疾病基因的预测。

Many disease genes are located within one or more specific chromosomal regions through position candidate approaches and genome-wide association studies. Prioritizing candidate genes by computational algorithms is important strategy to speed the identification of disease genes. Most approaches to identify disease genes based on function annotations have been presented in recent years. Most of them,starting from the function annotations of known genes associated with diseases,however,can not be used to identify genes for diseases without any known pathogenic genes or related function annotations. For such diseases, a new method is proposed to retrieve ralated gene functional information by mining biomedical literature and protein function annotation database. Thus, the genes for diseases lacking known causative genes also could be identified based on the gene function annotations mined.