Biostatistics/Bioinformatics Core

生物统计和生物信息学核心将在研究工作的计算和统计方面为软组织肉瘤中的孢子研究人员提供支持，包括临床试验、实验室实验、分子图谱和测序的设计和分析，以及综合基因组分析。在临床试验设计阶段，核心成员将与主要研究人员一起进行方案审查。在此审查的基础上，将提供议定书的统计部分，概述主要的科学目标、要研究的人口、主要和次要终点、实验设计、分析计划和以概率为依据的目标样本量。在试验结束时，将进行数据分析，以评估议定书中规定的主要和次要终点的结果。在临床前研究中，核心成员将协助制定实验设计，并在研究结束时分析和解释数据。该核心还将协助对人类肉瘤样本进行大规模的分子表征，包括分析RNA表达、microRNA表达、突变和DNA拷贝数变化，以确定肿瘤特征、药物反应和生存的预测因素。核心成员将使用双样本t检验和考克斯比例风险模型来确定信息丰富的特征。不同类型的分子数据将被整合并进行路径分析，以阐明有助于肿瘤发生和肿瘤表型的途径。风险预测工具将使用基于这些特征的COX比例风险模型来开发。

The Biostatistics & Bioinformatics Core will support investigators of the SPORE in Soft Tissue Sarcoma in the computational and statistical aspects of their research efforts, including the design and analysis of clinical trials, laboratory experiments, molecular profiling, and sequencing, as well as integrated genomic analyses. In the clinical trial design phase, a core member will conduct a protocol review with the principal investigator. Based on this review, a statistical section for the protocol will be provided, outlining major scientific objectives, population to be studied, primary and secondary endpoints, experimental design, analysis plans, and a targeted sample size justified in probabilistic terms. At the conclusion of the trial, data analyses will be performed to assess outcomes of the primary and secondary endpoints stated in the protocol. In preclinical studies, core members will assist in the formulation of the experimental design and in the analysis and interpretation of the data at the conclusion of the study. The core will also assist in the large-scale molecular characterization of human sarcoma samples, which involves analyzing RNA expression, microRNA expression, mutations, and DNA copy number changes to identify predictors of tumor characteristics, drug response, and survival. Core members will use two-sample t-tests and Cox proportional hazards models to identify informative features. The different types of molecular data will be integrated and pathway analysis performed to elucidate the pathways that contribute to tumorigenesis and tumor phenotype. Risk prediction tools will be developed using Cox proportional hazards models based on these features.