Discovery of Mental Health and Inflammation (MHAIN) Interactome

描述（由申请人提供）：蛋白质-蛋白质相互作用（ppi）的知识对于理解生物体的系统水平方面是必要的，包括精神过程。PPI网络（相互作用组）作为几种生物医学研究的载体。它可以用来了解疾病机制，药物靶点和副作用，以及疾病的遗传原因。人类有1014个基因与“大脑”相关；其中的448个基因甚至没有一个PPI是已知的。虽然发现哪些蛋白质对相互作用是有用的，但它也非常具有挑战性，因为超过99.9%的蛋白质对不相互作用。本工作的目的是开展系统设计的计算工作，以发现人类心理健康与炎症（MHAIN）相互作用组。MHAIN相互作用组指的是ppi网络，其中两种蛋白质中至少有一种与大脑或炎症有关。在发现新的蛋白质-蛋白质相互作用以构建相互作用组的过程中，将面临许多挑战。来自不同计算领域的成熟算法，如机器学习和信号处理，将被应用于实现所提出的目标。此外，神经心理过程和炎症过程相互影响的途径将从相互作用中挖掘出来。选定的相互作用将通过湿实验室实验进行验证。用于预测MHAIN相互作用组的方法是主动学习（获得将提供最大影响的蛋白质的标签和特征），迁移学习（将蛋白质特征和相互作用的知识从一个物种转移到另一个物种），多传感器融合（智能地整合来自不同预测器的输出）。通过作为精神健康生物医学研究的中心资源，预测的相互作用组将加速精神健康相关疾病的生物学和治疗的发现。对于每个感兴趣的蛋白质，将产生可验证的相互作用假设，将该蛋白质相互作用的搜索空间从大约25,000个减少到几个可能性。与重性抑郁症、阿尔茨海默病患者的精神病和全身性炎症（与精神疾病有着复杂的联系）有关的蛋白质的新ppi，将由专门研究这些领域的共同研究者放在生物医学的背景下。相互作用组将识别“中心蛋白”，这是许多途径的核心。它将提供新蛋白质（即那些目前一无所知的蛋白质）的“功能连接”假说。此外，许多疾病被发现无法通过抑制单一途径来治疗，并且通常需要操纵多个途径，而MHAIN相互作用组可以提供可能用于治疗疾病的多个途径的重叠点。因此，MHAIN相互作用组可以对我们对心理健康分子基础的理解产生广泛的影响，这将是拟议研究的结果。1

DESCRIPTION (provided by applicant): Knowledge of protein-protein interactions (PPIs) is necessary to understand system-level aspects of organisms including that of psychiatric processes. The PPI network (interactome) acts as a vehicle for several types of biomedical research. It can be used to understand the disease mechanisms, drug targets and side effects, and genetic causes for disease. There are 1,014 human genes associated with 'brain'; for 448 of these genes not even a single PPI is known today. While it is useful to discover which pairs of proteins interact, it is also exceptionally challenging as more than 99.9% of protein pairs do not interact. The objective of this work is to carry out systematically designed computational work to discover the human mental health and inflammation (MHAIN) interactome. The MHAIN interactome refers to the network of PPIs where at least one of the two proteins is involved in either brain or inflammation. Many challenges will be addressed in discovering new protein-protein interactions towards building the interactome. Well-established algorithms from diverse computational fields, such as machine learning and signal processing will be applied to achieve the proposed goals. Further, pathways of influence of neuropsychological processes and inflammatory processes on each other will be mined from the interactome. Selected interactions will be validated by wet-lab experiments. The approaches that will be employed for predicting the MHAIN interactome are proactive-learning (to obtain the labels and features of proteins that would provide the maximum impact), transfer-learning (to transfer the knowledge of protein features and interactions from one species to another), multi-sensor fusion (to intelligently integrate output from different predictors). The predicted interactome would accelerate discovery of the biology and treatment of mental health related diseases, by serving as a central resource for biomedical research on mental health. For every protein of interest, verifiable hypothesis of its interactions will be generated that reduce the search space of interactions of that protein from about 25,000 to a few possibilities. New PPIs of proteins involved in major depressive disorder, psychosis in patients with Alzherimer's disease, and systemic inflammation (which is intricately connected to psychiatric diseases), will be put in biomedical context by co- investigators who specialize in these areas. The interactome will identify "hub-proteins" that are central to many pathways. It will provide hypothesis of "functional connectivity" of new proteins (namely, those of which nothing is currently known). Further, many diseases are found to be untreatable by suppressing a single pathway, and often require manipulating multiple pathways, and the MHAIN interactome can provide the overlap points of multiple pathways that may be used for treatment of diseases. The MHAIN interactome, which would be the outcome of the proposed research, can thus have a broad impact on our understanding of molecular basis of mental health. 1