微生物群落是地球上生物多样性最丰富的资源，揭秘微生物群落对人类健康、自然环境及农业等领域意义重大。高通量测序为探寻微生物群落提供准确到核酸层面的有力数据。现有的研究主要基于微生物物种的基因组或16S rRNA参考序列，分析结果较为粗糙或不完整。基于序列特征的方法不依赖参考序列，但在高通量测序数据上的应用刚起步，其模型准确性、信息细致性和实际操作性有待完善。..本研究将构建高通量测序数据下微生物群落的高性能计算平台，基于不同的信息粒度逐步深入进行不同层面的生物分析。①不依赖参考序列，建立自适应统计模型衡量微生物群落的Beta多样性；②通过长特征序列识别不同类别微生物群落的特异物种/基因；③基于序列分布统计特性分析群落的物种结构组分；④在以上研究基础上，构建微生物群落比较、分析方法体系，设计后续的生物注释，建立开源、并行的计算平台，帮助研究同行全面理解人体及自然环境中的各类微生物群落。

Microbial communities have the richest biological diversity on the Earth. It is significant to reveal the microbial communities for the human healthy, environment and agriculture. High throughput sequencing technique offers accurate nucleotide sequences to explore the microbial communities. Most of the previous studies were alignment-based on reference sequences, which lead to the roughness and incompleteness of the analysis. The alignment-free methods on sequence signatures for high throughput sequencing data were initially developed , and their accuracy and feasibility are far from satisfactory. .. This proposal is going to build an efficient computational platform for analyzing microbial communities on high throughput sequencing data. Biological analysis of different levels are designed on different granularity. (1)Free from reference sequences, the adaptive statistical model to measure the Beta diversity of communities will be developed. (2)The specific species and genes will be identified based on the feature selection and classification on high-dimension sequence signatures. (3)The community profiling will be analyzed with semi-supervised clustering on statistical characteristics. (4)In the proposal, the methodology for comparison, analysis and understanding of microbial communities will be constructed, and the further biological annotations will be designed. The computational platform is going to give a comprehensive exploration and understanding of microbial communities.