HiComet: a high-throughput comet analysis tool for large-scale DNA damage assessment

AutoComet: A fully automated algorithm to quickly and accurately analyze comet assays

DNA damage is a common cellular feature seen in cancer and neurodegenerative disease, but fast and accurate methods for quantifying DNA damage are lacking. Comet assays are a biochemical tool to measure DNA damage based on the migration of broken DNA strands towards a positive electrode, which creates a quantifiable 'tail' behind the cell. However, a major limitation of this approach is the time needed for analysis of comets in the images with available open-source algorithms. The requirement for manual curation and the laborious pre- and post-processing steps can take hours to days. To overcome these limitations, we developed AutoComet, a new open-source algorithm for comet analysis that utilizes automated comet segmentation and quantification of tail parameters. AutoComet first segments and filters comets based on size and intensity and then filters out comets without a well-connected head and tail, which significantly increases segmentation accuracy. Because AutoComet is fully automated, it minimizes curator bias and is scalable, decreasing analysis time over ten-fold, to less than 3 s per comet. AutoComet successfully detected statistically significant differences in tail parameters between cells with and without induced DNA damage, and was more comparable to the results of manual curation than other open-source software analysis programs. We conclude that the AutoComet algorithm provides a fast, unbiased and accurate method to quantify DNA damage that avoids the inherent limitations of manual curation and will significantly improve the ability to detect DNA damage.

B-Comet Assay (Comet Assay on Buccal Cells) for the Evaluation of Primary DNA Damage in Human Biomonitoring Studies

Many subjects perceive venous blood collection as too invasive, and thus moving to better-accepted procedures for leukocytes collection might be crucial in human biomonitoring studies (e.g., biomonitoring of occupational or residential exposure to genotoxins) management. In this context, primary DNA damage was assessed in buccal lymphocytes (BLs), fresh whole venous, and capillary blood leukocytes, and compared with that in peripheral blood lymphocytes (PBLs)—the most frequently used cells—in 15 young subjects. Mouthwashes were collected after the volunteers rinsed their mouths with normal saline, and BLs were isolated by density gradient centrifugation. Blood samples were collected by venipuncture or by lancet. Anthropometric and lifestyle information was obtained by the administration of a structured questionnaire. As shown in the Bland-Altman plots, the level of agreement between BLs and PBLs lied within the accepted range, we thus enrolled a wider population (n = 54) to assess baseline DNA damage in BLs. In these cells, mean values of tail length (µm), tail intensity (%), and tail moment were 25.7 ± 0.9, 6.7 ± 0.4 and 1.0 ± 0.1, respectively. No significant association was observed between sex and smoking habit with any of the DNA damage parameters. Conversely, underweight subjects displayed significantly higher genomic instability compared with normal weight group (p < 0.05). In conclusion, we successfully managed to set up and update a non-invasive and well-accepted procedure for the isolation of BLs from saliva that could be useful in upcoming biomonitoring studies.

Deep learning method for comet segmentation and comet assay image analysis

Comet assay is a widely used method, especially in the field of genotoxicity, to quantify and measure DNA damage visually at the level of individual cells with high sensitivity and efficiency. Generally, computer programs are used to analyze comet assay output images following two main steps. First, each comet region must be located and segmented, and next, it is scored using common metrics (e.g., tail length and tail moment). Currently, most studies on comet assay image analysis have adopted hand-crafted features rather than the recent and effective deep learning (DL) methods. In this paper, however, we propose a DL-based baseline method, called DeepComet, for comet segmentation. Furthermore, we created a trainable and testable comet assay image dataset that contains 1037 comet assay images with 8271 manually annotated comet objects. From the comet segmentation test results with the proposed dataset, the DeepComet achieves high average precision (AP), which is an essential metric in image segmentation and detection tasks. A comparative analysis was performed between the DeepComet and the state-of-the-arts automatic comet segmentation programs on the dataset. Besides, we found that the DeepComet records high correlations with a commercial comet analysis tool, which suggests that the DeepComet is suitable for practical application.