PDF(Pubmed)
Abstract：
Recognizing the \"essential\" factors that contribute to a clinical outcome is critical for designing appropriate therapies and prioritizing limited medical resources. Demonstrating a high correlation between a factor and an outcome does not necessarily imply an essential role of the factor to the outcome. Human protective adaptive immune responses to pathogens vary among (and perhaps within) pathogenic strains, human individual hosts, and in response to other factors. Which of these has an \"essential\" role? We offer three statistical approaches that predict the presence of newly contributing factor(s) and then quantify the influence of host, pathogen, and the new factors on immune responses. We illustrate these approaches using previous data from the protective adaptive immune response (cellular and humoral) by human hosts to various strains of the same pathogenic bacterial species. Taylor\'s law predicts the existence of other factors potentially contributing to the human protective adaptive immune response in addition to inter-individual host and intra-bacterial species inter-strain variability. A mixed linear model measures the relative contribution of the known variables, individual human hosts and bacterial strains, and estimates the summed contributions of the newly predicted but unknown factors to the combined adaptive immune response. A principal component analysis predicts the presence of sub-variables (currently not defined) within bacterial strains and individuals that may contribute to the combined immune response. These observations have statistical, biological, clinical, and therapeutic implications.
摘要：
认识到有助于临床结果的“基本”因素对于设计适当的疗法和优先考虑有限的医疗资源至关重要。证明因素与结果之间的高度相关性并不一定意味着因素对结果的重要作用。人类对病原体的保护性适应性免疫反应在病原菌株之间（也许在内部）有所不同。人类个体宿主，和其他因素的反应。我们提供了三种统计方法来预测新贡献因子的存在，然后量化宿主的影响，病原体,以及免疫反应的新因素。我们使用人类宿主对相同病原细菌物种的各种菌株的保护性适应性免疫应答（细胞和体液）的先前数据来说明这些方法。泰勒定律预测，除了个体宿主间和细菌内物种菌株间变异性外，还存在其他可能导致人类保护性适应性免疫反应的因素。混合线性模型测量已知变量的相对贡献，个体人类宿主和细菌菌株，并估计新预测但未知因素对联合适应性免疫反应的总贡献。主成分分析预测可能有助于组合免疫应答的细菌菌株和个体内的子变量(目前未定义)的存在。这些观察有统计学意义，生物，临床,和治疗意义。
