The need for standardizing comet assay image analysis: a comparison of comet assay scoring software

The comet assay is a simple method to assess DNA damage. Although standardized protocols exist, there is no standard protocol for comet image analysis, representing a potential source of inter-laboratory variability. Several scoring programs exist, including commercially available (e.g. Comet IV) and freeware (e.g. CaspLab, CometScore and OpenComet). As costs frequently drive the selection of freeware, it is essential to evaluate their performance when compared to Comet IV (considered the reference method). This study compared these programs using comet images with varying levels of DNA damage, categorized into three data sets: (A) the full data set comprising comets with an extensive range of % Tail DNA, (B) a subset comprising comets with <15% Tail DNA (low damage), and (C) a subset with >50% Tail DNA (high damage). Across the full dataset and at high levels of DNA damage, most freeware programs differed significantly from Comet IV. CometScore was the only freeware that did not significantly differ from Comet IV in all three subsets. Correlation and agreement analyses revealed weaker associations for subset B, especially for OpenComet. CaspLab and OpenComet exhibited systematic and/or proportional biases, whereas CometScore consistently showed strong correlations and minimal bias relative to CometIV. Collectively, these findings demonstrate that freeware programs can provide acceptable performance, but they may deviate from Comet IV. Our results highlight the importance of informed software selection and the need for standardized and validated image analysis workflows, to improve the accuracy, reproducibility, and inter-laboratory comparability of comet assay data.