Addendum: The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity

Addendum to: Nature https://doi.org/10.1038/nature11003, published online 28 March 2012; addendum 28 November 2012.

The Supplementary Information of our original Letter (https://doi.org/10.1038/nature11003) and the Addendum published on 28 November 2012 (https://doi.org/10.1038/nature11735; hereafter referred to as ‘first Addendum’) contain all of the data used for our analyses. These data include normalized activity (growth inhibition) responses for every cell line at each dose for all 24 of the compounds and drugs tested. These activity values were calculated from replicated data points extracted from one or more experimental assay plates, and normalized using treated and untreated controls from the same plates; as detailed in the original Letter (see the original Supplementary Information 2, page 10), and in the first Addendum. In subsequent correspondence, readers expressed an interest in reproducing our experimental pre-processing, so we shared with them the raw Cell Titer Glo (CTG) data values from each experimental well. These data are now provided as a reference for future readers.

Supplementary Data 1 to this Addendum contains raw values from wells with cell lines growing under control (AC, positive active; NC, untreated negative) or drug-treated (SA, single-agent) conditions. Supplementary Data 1 contains two row entries per well (one with raw CTG readouts, and the other with calculated activity values), and each column contains descriptions or response values, as detailed in Table 1 of this Addendum. Only valid wells are included, so there may be missing concentrations from curves containing data that were flagged as ‘invalid’. Compound and cell line names in Supplementary Data 1 are as they were originally captured, which in some cases may differ from names used in the Supplementary Tables of the original Letter and first Addendum. Calculated activity values may also differ in some cases from those in the original Letter, owing to subsequent quality-control call changes in our database since 2012. Chemical names can be reconciled using Table 2 of this Addendum, noting that ‘sample-ID’ denotes the ‘ID-concept’ of Table 2 with salt form and batch suffixes appended. Discrepant cell line names may be understood with reference to columns A, C and D of the ‘All_Drug_response_data’ sheet in supplementary data 1 from the 2015 CCLE and GDSC consortium paper¹.

We thank Jessie Ambrose and Vladimir Baranov for extracting and quality-checking the raw chemical response data in this Addendum. We also thank the reader who requested the raw data and assisted in quality-checking these results.

Table 1 Descriptions for key columns in Supplementary Data 1

Table 2 Chemical name mappings