基于多层次微进化特征的自然选择压力检测研究

自然选择压力检测是分子进化领域重要的基础研究内容之一，且微进化特征是分子进化模型和自然选择压力检测算法的核心。然而，现有选择压力检测算法所考虑的微进化特征各不相同，造成其检测结果也迥然不同，严重影响分子进化等相关研究，其根本原因在于长期缺乏对微进化特征的系统整合分析，缺少基于多层次微进化特征的选择压力检测算法。因此，本项目拟从以上角度出发，系统整合分析分子序列不同层次的多个微进化特征，挖掘提取分子进化的主效特征，建立基于多层次微进化特征的分子进化新模型，结合考虑多位点同时突变和序列区域异质性进化，发展基于多层次微进化特征的选择压力检测新算法和新工具，采用最大似然法设计实现自然选择压力的准确计算和异质性进化区域的无监督聚类，为分子进化、基因功能、生物适应性演化机制等研究提供新思路和新方法。

Detection of natural selection acting on molecular sequences is of fundamental significance in molecular evolution. Microevolution features (MEF) obtained from molecular sequences are critical in modeling sequence evolution and detecting natural selection. Existing algorithms factored multiple different MEFs into estimating natural selection, accordingly leading to different estimates of natural selection and affecting the reliability of studies on molecular evolution. However, very few attention has been devoted to systematic integrative analyses on MEFs, not mention to detect natural selection based on principal MEF. To address these issues, here we propose to perform systematic integrative analyses on MEFs, extract principal features from all MEFs and devise a new evolutionary model that factors principal MEFs, allows multiple-nucleotide simultaneous mutations and accounts for heterogeneous selective pressures on different clustered regions. Based on principal MEFs and implemented in a maximum-likelihood framework, we aim to develop a new algorithm as well as a new tool that is able to yield accurate estimation of natural selection and achieve unsupervised identification of clusters with heterogeneous selective pressures, which accordingly would provide new insights and methods for in-depth investigation of molecular sequence evolution, gene function and adaptive evolution mechanisms.