非同义单核苷酸变异致病性预测研究综述

Review on pathogenicity prediction studies of non-synonymous single nucleotide variations

超过6000种人类疾病是由非同义单核苷酸变异(Non-synonymous single nucleotide variations,nsSNVs)引发的,快速准确地预测nsSNVs的致病性,有助于理解发病原理和设计新药物,也是生物信息领域的重要研究课题之一。该文给出了nsSNVs致病性研究的重要意义与背景知识;总结了国内外研究的主流方法,包括基于突变频率的方法、基于通路的方法、结合基因组和转录信息的方法、基于序列进化保守性的方法、基于序列和结构混合特征的方法以及综合评价类方法,对代表性方法进行了阐述;给出了nsSNVs致病性研究中常用的数据库、特征表示方法以及性能评价指标,多角度地对12种nsSNVs致病性预测方法进行了比较;最后,展望了nsSNVs致病性预测中可能取得突破的若干研究方向。

More than six thousand human diseases are caused by non-synonymous single nucleotide variations(nsSNVs).Fast and accurate prediction of the pathogenicity of nsSNVs helps to understand the pathogenesis and to design new drugs and is one of the extremely important research issues in the field of bioinformatics.This paper gives the significance and relevant background knowledge in nsSNVs pathogenicity research.Current research methods are also summarized here,including mutation frequency-based methods,pathway-based methods,methods combined with genomes and transcriptomes information,methods based on sequence evolutionary conservation,methods based on the mixed features of protein sequence and structure,and comprehensive evaluation methods.Some representative tools are also expounded here.This paper summarizes several databases,feature representation methods and evaluation indexes,which are commonly utilized in the study of nsSNVs.This paper also compares twelve methods of pathogenicity of nsSNVs prediction from multiple perspectives.Some research directions that can make breakthroughs in the prediction of pathogenicity of nsSNVs are preticted here.