基于多信息融合的酵母重组冷热点预测  

Predicting recombination hot/cold spots in Saccharomyces cerevisiae from multiple information

作　　者：张冰洁[1] 刘国庆[1,2] 蔡禄[1,2] 

机构地区：[1]内蒙古科技大学数理与生物工程学院,包头014010 [2]内蒙古科技大学生物工程与技术研究所 [3],包头014010

出　　处：《科学通报》2014年第11期953-959,共7页Chinese Science Bulletin

基　　金：国家自然科学基金(61102162,61271448)资助

摘　　要：减数分裂重组并非均匀发生在基因组上,而是在一些区域有着较高的重组频率(重组热点),在另一些区域重组频率较低(重组冷点).重组的发生不仅与序列特征有关,还依赖于染色质的结构.准确鉴定重组热点和冷点对于认识重组发生的分子机制以及基因组进化规律具有重要意义.本文首先在实验数据的基础上识别2 kb尺度的重组冷热点,然后采用多样性增量结合二次判别分析(IDQD)和支持向量机(SVM)算法,基于一系列与DNA序列、结构及其热力学稳定性以及染色质结构相关的特征对酵母的重组冷热点进行了分类预测.结果表明,预测模型能够有效区分重组冷热点;从预测结果的敏感性、特异性和总精度来看,IDQD算法优于SVM算法.Meiotic recombination is non-uniformly distributed across the genome, occurring at relatively high frequencies in some genomic regions （hotspots） and at relatively low frequencies in others （coldspots）. Recombination depends not only on sequence features, but also on chromatin structure. Accurate identification of recombination hotspots and coldspots would assist our understanding of genetic recombination mechanisms and genome evolution. In this study, 2 kb recombination hot/cold spots were identified in Saccharomyces cerevisiae from experimental data. Using increment diversity with quadratic discriminant （IDQD） and support vector machine （SVM） algorithms, we predicted hot/cold spots from multiple genomic features associated with sequence, structure, thermodynamic stability of DNA and chromatin structure. The predictive model accurately discriminated recombination hotspots from coldspots. IDQD yielded higher overall performance than SVM in terms of sensitivity, specificity and accuracy.

关 键 词：重组热点 k-mer频数 热力学稳定性 核小体占据率 

分 类 号：Q93[生物学—微生物学]