Reactome: An Open Knowledgebase of Human Pathways

项目摘要 我们寻求更新反应组人类生物知识库的核心运营资金 途径和过程。Reactome是一个经过管理的、开放访问的生物分子途径数据库，它可以自由地 由生物研究界的所有成员使用和重新分发。它是临床医生使用的，吉内蒂- 科学家、基因组学研究人员和分子生物学家解释高通量实验的结果 生物信息学家寻求开发从基因组研究中挖掘知识的新算法， 以及系统生物学家建立正常和疾病变异途径的预测模型。 我们的策展人，具有细胞和分子生物学背景的博士级科学家，与In- 社区内依赖调查人员汇编机器可读的人类生物学描述 小路。每条路径在发表之前都经过了广泛的检查和同行审查，以确保其断言 都得到了原始文献的支持，而且人类分子事件是从 动物模型有一个可审计的推理链。精选反应组途径目前覆盖8930蛋白质- 编码基因(占基因组翻译部分的44%)和~150个RNA基因。我们还提供一个网络 通过保守的机器学习方法预测可靠的‘功能交互’(FIS)，该方法涵盖了 额外的3300个基因，覆盖了大约60%的已知基因组。 在接下来的五年里，我们将：(1)策划新的大分子实体，具有临床意义的蛋白质 序列变异体和异构体，类药物分子，以及这些实体形成的络合物，形成新的反应- (2)用针对重大疾病的替代途径来补充正常的途径，并发展- 操作生物学；(3)扩展和自动化我们的管理、管理和社区注释工具； (4)使用概率图模型和布尔网络集成路径建模技术 路径和网络扰动研究；(5)开发针对以下目标的更多引人注目的软件界面 计算机和实验室生物学家用户；以及(6)并改善与生物信息学家的联系，分子双 科学家和临床研究人员。

Project Summary We seek renewal of the core operating funding for the Reactome Knowledgebase of Human Biological Pathways and Processes. Reactome is a curated, open access biomolecular pathway database that can be freely used and redistributed by all members of the biological research community. It is used by clinicians, geneti- cists, genomics researchers, and molecular biologists to interpret the results of high-throughput experimental studies, by bioinformaticians seeking to develop novel algorithms for mining knowledge from genomic studies, and by systems biologists building predictive models of normal and disease variant pathways. Our curators, PhD-level scientists with backgrounds in cell and molecular biology work closely with in- dependent investigators within the community to assemble machine-readable descriptions of human biological pathways. Each pathway is extensively checked and peer-reviewed prior to publication to ensure its assertions are backed up by the primary literature, and that human molecular events inferred from orthologous ones in animal models have an auditable inference chain. Curated Reactome pathways currently cover 8930 protein- coding genes (44% of the translated portion of the genome) and ~150 RNA genes. We also offer a network of reliable ‘functional interactions’ (FIs) predicted by a conservative machine-learning approach, which covers an additional 3300 genes, for a combined coverage of roughly 60% of the known genome. Over the next five years, we will: (1) curate new macromolecular entities, clinically significant protein sequence variants and isoforms, and drug-like molecules, and the complexes these entities form, into new reac- tions; (2) supplement normal pathways with alternative pathways targeted to significant diseases and devel- opmental biology; (3) expand and automate our tools for curation, management and community annotation; (4) integrate pathway modeling technologies using probabilistic graphical models and Boolean networks for pathway and network perturbation studies; (5) develop additional compelling software interfaces directed at both computational and lab biologist users; and (6) and improve outreach to bioinformaticians, molecular bi- ologists and clinical researchers.