Understanding cascading cellular protein responses following multi-protein stimuli using network modeling and real-world evidence

项目摘要/摘要：据估计，药物不良事件占医疗保健的16% 在美国，只有不到10%的新疗法进入临床，这表明我们需要更好的模型 因为他预见到了毒品的影响。有趣的是，细胞对刺激的反应是多路的，并且是级联效应， 比如药物疗法。例如，一种药物刺激可以导致多种结果，包括改变疾病 症状或引起不良副作用，此外，药物可以影响其下游的蛋白质 预定目标。然而，药物疗法并不是按常规设计的，它们具有多重的、级联的效应或多重的 考虑蛋白质结合特性，因为我们缺乏了解这些特性的定量机制模型 有趣的细胞效应。蛋白质相互作用(PPI)网络模型已经确定了下游蛋白质 与疾病和与药物治疗相关的副作用有关的，展示了将多个 刺激的结果。这些关联模型不足以确定药物靶蛋白的优先顺序，因为 PPI网络缺乏机制细节来描述细胞反应的大小或相对贡献 多蛋白质刺激。这项研究的总体目标是推导出多个变量之间的定量关系 使用临床数据的蛋白刺激和下游反应蛋白。这项建议是创新的，因为 上下文特定交互(CSI)方法：与其他PPI网络方法相比，CSI分析 在以下情况下，预测准确率和精确度分别提高了50%和76-95% 使用药物靶标下游的PPI识别严重不良事件并强调临床数据 整合。这种方法强调了使用表型学习PPI网络参数的重要性。 具体数据，以更好地了解所有与网络相关的表型，并论证推导的可行性 PPI网络模型中的机械性细节。这一计划意义重大，因为它将减少总体副作用- 通过促进对细胞多重反应的更好理解来影响和提高治疗效果 对多种蛋白质的刺激。此外，我还展示了始终如一的工作效率以及 从事研究项目，建立卓越的独立事业。

Project Summary/ Abstract: Drug adverse events have been estimated to contribute to 16% of healthcare spending in the US, and less than 10% of new treatments reach the clinic, suggesting we need better models for anticipating drug effects. Intriguingly, cells have multiplexed, and cascading effects in response to stimuli, such as drug therapies. For instance, a drug stimulus can induce multiple outcomes including modifying disease symptoms or causing undesirable side-effects, and further, drugs can influence proteins downstream of their intended targets. Yet, drug therapies are not routinely designed with their multiplexed, cascading effects or multi- protein binding properties in mind because we lack quantitative mechanistic models for understanding these interesting cellular effects. Protein-protein interaction (PPI) network models have identified downstream proteins associated with diseases and side-effects relevant to drug therapies, demonstrating the ability to link multiplexed outcomes with a stimulus. These associative models are insufficient for prioritizing drug target proteins because PPI networks lack mechanistic detail to describe the magnitude or relative contribution of cellular responses to multi-protein stimuli. The overall objective of this research is to derive quantitative relationships between multi- protein stimuli and downstream response proteins using clinical data. This proposal is innovative because of the context-specific interaction (CSI) approach: compared to other PPI network approaches, CSI analysis demonstrated a 50% and 76-95% improvement in prediction accuracy and precision, respectively, when identifying severe adverse events using PPIs downstream of drug targets and the emphasis on clinical data integration. This approach emphasized the importance of learning PPI network parameters using phenotype- specific data to better understand all network-associated phenotypes and demonstrated the feasibility of deriving mechanistic details in PPI network models. This program is significant because it stands to reduce overall side- effects and increase therapeutic efficacy by advancing a better understanding of cellular multiplexed responses to multi-protein stimuli. Further, I have demonstrated consistent productivity and the flexibility and adaptability in pursuing research projects to establish a distinguished independent career.