Assessing the influence of different alignment tools on the accuracy of a forensic epigenetic clock.

Abstract

DNA methylation (DNAm) has long been a commonly investigated biomarker in biomedical research. The current gold standard for DNAm detection is bisulfite sequencing which requires dedicated alignment tools that can handle reduced sequence complexity. One commonly used application of DNAm are epigenetic clock measurements. These clocks have been adapted by many fields for their specific needs, including forensic genetics. Here, epigenetic clocks were designed to help estimate the chronological age of a biological stain donor for investigative purposes. In this study, data generated with a well established forensic epigenetic clock is aligned with four different bisulfite-specific alignment tools: 'Bwa-meth', 'Abismal', 'Bismark', and 'BS-Seeker2'. For each tool, we tested up to six different settings, altering parameters such as the maximum number of mismatches or the score function setting. The goal was to investigate whether the final predicted ages differed considerably between the tested alignment tools and settings. Quality controls such as read depth, precision, recall, F1 score, and alignment run time were also assessed. To allow other researchers to easily perform such methylation comparison analyses on their own data, a Shiny app called 'MethylAge Explorer' was developed within this study. None of the tested settings for the three alignment tools 'Abismal', 'Bismark', and 'BS-Seeker2' outperformed the originally used alignment tool 'Bwa-meth' in terms of age prediction accuracy. However, differences in final age predictions were observed between the different alignment tools. Therefore, it is necessary to be aware of which alignment tool to use for particular epigenetic clocks. The data underlying this article and the code for the shiny app are available on GitHub (https://github.com/charlsut/methylage_explorer). Supplementary data are available at Bioinformatics online.