作者: Bin Li , Hyunjin Shin , Georgy Gulbekyan , Olga Pustovalova , Yuri Nikolsky
DOI: 10.1371/JOURNAL.PONE.0130700
关键词: Gene regulatory network 、 Medicine 、 Erlotinib Hydrochloride 、 Clinical study design 、 Drug 、 Clinical trial 、 Sorafenib 、 Computational biology 、 Erlotinib 、 Bioinformatics 、 Drug discovery 、 General Biochemistry, Genetics and Molecular Biology 、 General Agricultural and Biological Sciences 、 General Medicine
摘要: Development of drug responsive biomarkers from pre-clinical data is a critical step in discovery, as it enables patient stratification clinical trial design. Such translational can be validated early phases and utilized inclusion parameter later stage trials. Here we present study on building accurate selective sensitivity models for Erlotinib or Sorafenib vitro data, followed by validation individual corresponding treatment arms generated the BATTLE trial. A Partial Least Squares Regression (PLSR) based modeling framework was designed implemented, using special splitting strategy canonical pathways to capture robust information model building. predictive could used identify sub-group patients that respond better treatment, these are specific drugs. The derived signature genes reflect each drug’s known mechanism action. Also, predict potential cancer indications consistent with results selection globally normalized GEO expression datasets.