Approaches for the Exploitation of Data Derived from Pharmacogenomics Screens

Since the discovery of the first immortal cell line HeLa in 1951, cancer cell lines (CCLs) have been a mainstay of cancer research. But it was not until the establishment of the National Cancer Institute of 60 cancer cell lines (NCI-60) that they became invaluable to cancer drug development. Over 30 years on from the beginning of large-scale cancer drug screening, many large-scale CCL panels now exist, including the Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC). These studies have had the reproducibility of their results questioned. With the increased attention of the reproducibility crisis and previous issues with cell line misidentification, scepticism is perhaps fair. It is determined in analyses here that platinum drugs do not all have the same drug sensitivity profiles and confirmed a novel approach to determine cancer cell line sensitivity using clinically relevant thresholds. Data from two large-scale pharmacogenomics screens was also used to validate experimental work carried out with the compound YM155 with relation to ABCB1 and SLC35F2 expression levels. Reproducibility and replication of results is also examined using NCI-60 data and it was determined that high variability in drug sensitivity data existed despite standardisation of protocols. Results here prove that if used correctly, despite flaws in the panels, the data can prove insightful in many ways. Particularly when used as a validation method for research carried out in a wet lab. It was also shown that reproducibility of data from these CCL panels should be considered when using for analysis as drug sensitivity data was shown to be highly variable, and that perhaps, despite high levels of standardisation there exists an intrinsic variability in biological systems.