Statistical and computational model for high dimensional data analysis

高维数据分析的统计和计算模型

• 批准号: RGPIN-2016-06546
• 负责人: Wong, WilliamWaiLun
• 金额: $ 1.31万
• 依托单位: University of Waterloo
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2020
• 资助国家: 加拿大
• 起止时间: 2020-01-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=709663
• 关键词: Statistical; computational; model; dimensional; data

Research towards the development of statistical and computational methodologies for analyzing high dimensional data has gained much attention over the last decade. Methodological advances with respect to feature selection, classification, and predication have been widely applied in many fields, such as the drug discovery process, high throughput genomic and genetic data analysis, and infectious disease modeling. A drug is often a small molecule that interacts with the binding sites of some target proteins. The drug discovery process can typically be divided into four main stages: (i) target identification; (ii) lead discovery; (iii) clinical trials; and (iv) regulatory approval and reimbursement. Drug discovery is a time consuming and costly process. Computer-Aided Drug Design (CADD) is a specialized discipline that uses computational statistics models and algorithms to aid the drug discovery process. The long-term objective of my NSERC DG research program is to develop effective and efficient statistical and computational methods for analyzing diverse types of high dimensional data in the CADD areas. In the next 5 years, my short-term objectives are to make advances in three main themes: 1) computational drug discovery: from pharmacoinformatics to pharmacoeconomics via multi-task statistical learning; 2) statistical methods for genetic association studies; and 3) individual-level modeling of infectious diseases in large populations. The proposed research provides new strategies for CADD and will have direct impact on the planning stage of the research and development of new drugs. New territory in the research area of CADD will be explored by using pharmacoeconomics to influence the direction of drug discovery via multi-task statistical learning, detecting gene-gene interactions via genome-wide haplotype-based association analysis, and improving agent-based modeling calibration process by using the approximate Bayesian computation method.

在过去的十年里，对高维数据分析的统计和计算方法的发展的研究已经得到了很大的关注。在特征选择、分类和预测方面的方法学进展已经广泛应用于许多领域，例如药物发现过程、高通量基因组和遗传数据分析以及传染病建模。 药物通常是与某些靶蛋白的结合位点相互作用的小分子。药物发现过程通常可以分为四个主要阶段：（i）目标识别;（ii）先导发现;（iii）临床试验;以及（iv）监管批准和报销。药物发现是一个耗时且昂贵的过程。计算机辅助药物设计（CADD）是一门专业学科，它使用计算统计模型和算法来帮助药物发现过程。 我的NSERC DG研究计划的长期目标是开发有效和高效的统计和计算方法，用于分析CADD领域的各种类型的高维数据。在接下来的5年里，我的短期目标是在三个主题上取得进展：1）计算药物发现：通过多任务统计学习从药物信息学到药物经济学; 2）遗传关联研究的统计方法; 3）大人群中传染病的个体水平建模。 该研究为CADD提供了新的策略，并将对新药研发的规划阶段产生直接影响。CADD研究领域的新领域将通过使用药物经济学通过多任务统计学习影响药物发现的方向，通过全基因组单倍型关联分析检测基因-基因相互作用，并通过使用近似贝叶斯计算方法改进基于代理的建模校准过程来探索。