Unsupported claim of significant discrimination between monozygotic twins from multiple pairs based on three age-related DNA methylation markers

Dear Editor
We highly welcome the development of innovative approaches that help differentiate between monozygotic (MZ) twins based on DNA evidence in forensic casework, given that standard forensic STR profiling does not allow their individual identification. However, while Marqueta-Gracia et al. recently claimed in the Journal that almost half of the tested MZ twin pairs were discriminated based on the same three age-related CpG fragments, we present arguments and evidence demonstrating that this claim is not supported by their data.
First, given the same chronological age of MZ twins per pair, applying age-related DNA methylation markers for discriminating MZ individuals within pairs appears counterintuitive. [...]
Second, any data-driven claim on individual discrimination depends on the magnitude of technical error of the method used for data generation, which is especially crucial if the observed individual differences are small. [...]
Third, Marqueta-Gracia et al. applied an approach to test for statistically significant inter-individual differences that appears inappropriate for the given purpose for several reasons. [...]
Fourth, it appears peculiar that, although the authors have analyzed a total of six age-related CpG fragments, they based their final conclusions regarding MZ twin discrimination on only three of them (ITGA2B, ASPA, and ZIC5). [...]
Fifth, cell type composition is a major challenge in epigenetic analysis. Differences in DNA methylation patterns between different cell types are well-known and typically require correction in epigenomewide association studies, an approach that should also be followed in targeted DNA methylation analysis. [...]
In summary, convincing data evidence that i) age-related CpGs allow for the reliable discrimination of MZ twin individuals within pairs, ii) a small set of the same CpGs allows for reliable MZ twin discrimination across different pairs, and iii) a small set of the same CpGs allows for reliable MZ twin discrimination from forensic trace-type materials, all remain to be presented.