Computational Biology Approaches to Support Biomarker Discovery and Development
Бин Ли, Hyunjin Shin, William L. Trepicchio, Andrew J. Dorner · 2020
Biomarkers based on disease pathways, the effects of target engagement, and patient heterogeneity are critical components of drug development. This chapter introduces computational biology approaches to support translational biomarker research for discovery and development in three key areas. They are text mining–derived diseases/drug maps; predictive modeling identifying and validating translational biomarkers for patient stratification; and building translational research platforms to integrate clinical and omics data, with predictive modeling. The chapter highlights a case study identifying and validating a translational biomarker for patient stratification and disease indication selection. The architecture of translational data repositories will be critical in the management and analysis of combinations of genomic, phenotypic, and clinical data for translational research. The cell line–derived erlotinib and sorafenib sensitivity models predicted BATTLE clinical trial progression-free survival outcomes with accuracies of 84% and 79%, respectively.