太空舱内微生物种群结构及功能分析相关计算方法

Microbes play a key role in maintaining life on earth, fixing gases and bringing the cycling of nutrients and compounds, both essential for the survival of all organisms. Without microbes, we couldn't eat or breathe. Facing high-throughput microbial gene sequence data generated by the next-generation sequencing (NGS), current methods and tools have a great neck-bottle for collecting, restoring and dealing with these microbial data. There is urgent need to develop novel theory and computational methods for analyzing these data, which will not only expand our understanding on microbial diversity, structure and function of microbial population, bacterial population distribution and their evolution in spacecraft, but also provide the technical support for keeping the astronaut health in a good condition and establishing the astronaut-spacecraft safeguard system. However, the microbial diversity, function and structure of microbial population in spacecraft are still poorly understood, historically due to most researchers adopted the cultivate technologies which just provide access to only a very small fraction of the microbial diversity and more than 99% of naturally occurring microbes are considered ‘unculturable’ on standard culture media. Even if a few researchers employed uncultured technology (e.g. matagenomic), they just used simply statistical approaches to analyze the microbial data, which provide limited information for our understanding microbial diversity and function. This project will address important computational challenges in analyzing microbial sequence data from the space vehicle simulator or spacecraft. According to national major strategic needs of manned space flight, we will collect the environment samples from space vehicle simulator or spacecraft at different time slot, generate microbial sequence data with NGS technology, develop effective microbial operational taxonomic unit (OTU) clustering algorithms, universal metagenomic gene prediction algorithms, more robust methods of metabolic pathway (or network) reconstruction with probability measure and metabolic network alignment algorithms for more samples. All above algorithms and methods developed can help to find the microbial diversity, population structure, function and evolution information in space vehicle simulator or spacecraft. In the end, we will develop and build software and pipelines to analyze the microbial sequence data from the space vehicle simulator and spacecraft, which will provide the technical reserves for our national deep space exploration in future.

微生物与人类生活密切相关，面对宏基因组学产生的海量微生物基因序列数据，现有信息收集、贮存、分析方法及工具远不能满足研究需求。发展新的理论和计算方法具有重要意义，不仅有助于了解太空密封环境内微生物群落结构、功能、菌群分布及演变规律，且有助于保障与改善航天员健康，建立人-机环境系统安全综合保障体系。当前太空舱及空间站等特定密闭环境内微生物研究大多基于微生物纯培养技术，且数据分析多采用简单的统计学方法，获得的信息量非常有限。鉴于此，本项目以国家载人航天重大战略需求为牵引，采用二代测序技术对太空模拟舱/太空舱内的微生物测序，研究有效的微生物操作分类单元（OTU）聚类算法、普适的宏基因组基因预测算法、高精度代谢途径/网络重构及比对算法，从海量的微生物序列数据中发现太空密封环境内微生物种群结构、功能、菌群分布及演变规律，并构建微生物种群结构、功能及菌群演化分析平台，为国家未来的深空探测提供技术储备。

微生物与人类生活密切相关，面对宏基因组学产生的海量微生物基因序列数据，现有信息收集、贮存、分析方法及工具远不能满足研究需求。发展新的理论和计算方法具有重要意义，不仅有助于了解太空密封环境内微生物群落结构、功能、菌群分布及演变规律，且有助于保障与改善航天员健康，建立人-机环境系统安全综合保障体系。本项目从密封环境下的微生物16S rRNA序列出发，严格按照项目计划书要求开展研究，并在RNA甲基化、复杂网络可控性及药物-靶点等方面也开展了研究工作，主要取得了以下研究成果：1、分别采集33天、180天密封环境下的人体部位、仪器表面及空气微生物样本，然后进行16S rRNA测序；基于这些16S rRNA序列，分析了密封环境中的微生物物种组成及其多样性。2、针对目前启发式OTU聚类算法对种子序列选取敏感、且不考虑测序误差问题，先后提出基于DB图的16S rRNA 序列启发式OTU聚类算法（DBH）和基于稠密子团与模块度挖掘的OTU聚类算法（DMclust），有效揭示微生物的物种多样性。3、针对目前代谢通量预测算法的基因表达阈值选取影响算法通用性问题，提出一种基于代谢通量最小化和Huber惩罚函数的通用代谢通量预测算法（HPCOF），并用于构建人类组织代谢网络。4、针对目前基因调控网络构建算法时间复杂度大、假阳性边较多问题，先后提出基于KNN局部贝叶斯模型的基因调控网络构建算法（LBN）和基于有序条件互信息与有限父结点的基因调控网络构建算法（OCMIPN），高精度、快速构建大规模基因调控网络。5、基于深度学习，分别提出宏基因组基因预测算法（Meta-MFDL）和长非编码RNA预测算法（lncRNA-MFDL），高精度预测宏基因片段序列及长非编码RNA序列。6、提出m6A-Driver驱动基因识别算法及CTC网络目标控制策略，有效识别m6A甲基化驱动基因及关键驱动基因集合。7、提出一种通用网络社团检测算法（CDMIC），有效检测无权/加权网络、无向/有向网络及符号网路社团结构。8、提出FET-HMM、DRME、QNB差异甲基化检测算法，小尺度、小样本下高精度分析RNA差异甲基化状态，发现共甲基化模式。9、提出NMCOM、AFMFSC致病基因预测算法，成功预测前列腺、肺癌潜在风险致病基因。10、提出药物-靶点作用预测算法，高精度为药物筛选靶点。