Cover Picture: How Consistent are Publicly Reported Cytotoxicity Data? Large‐Scale Statistical Analysis of the Concordance of Public Independent Cytotoxicity Measurements (ChemMedChem 1/2016)
Isidro Cortés‐Ciriano, Andreas Bender · ChemMedChem · 2016
The front cover picture shows a density plot for 50 959 pairs of pIC50 measurements from a stringently curated cytotoxicity data set. Blue signifies higher density, meaning high concordance between pairs of measurements. By analyzing public cytotoxicity data from ChEMBL v.19, a large-scale investigation into the reliability of this type of data was performed. Results show that data vary significantly between reports, even for the same cell type and compound system, and as a consequence, models for predicting cytotoxicity trained using publicly available data have inherent inaccuracy. Annotation errors are responsible for the high discordance observed for some pairs, pointing out the importance of data curation when automatically extracting cytotoxicity data from public databases. From this analysis, guidelines can be created on the reliability of cytotoxicity data from public databases, which could prove valuable for modeling purposes and guide in the reporting of this data type. More information can be found in the Full Paper by Isidro Cortés-Ciriano and Andreas Bender on page 57 in Issue 1, 2016 (DOI: 10.1002/cmdc.201500424).