Ghosh Debashis, Poisson Laila M
Department of Statistics, College of Medicine, Pennsylvania State University, University Park, PA 16802, USA.
Genomics. 2009 Jan;93(1):13-6. doi: 10.1016/j.ygeno.2008.07.006. Epub 2008 Sep 14.
With the development of new technologies for assaying biological activity on a global basis in experimental samples, various new "-omics" signatures have been developed to predict disease progression. Such signatures hold the potential to alter the nature of clinical management of human disease. In this article, we describe some necessary statistical considerations needed to take these signatures from the discovery phase to a clinically useful assay. Much of the work discussed is in the area of cancer.
随着在实验样本中对生物活性进行全球范围内检测的新技术的发展,已经开发出各种新的“组学”特征来预测疾病进展。这些特征有可能改变人类疾病临床管理的性质。在本文中,我们描述了将这些特征从发现阶段转化为临床可用检测方法所需的一些必要统计考量。所讨论的大部分工作都在癌症领域。