Meta Repeat Finder: Combining multiple detectors for tandem repeat discovery in protein sequences

Tandem repeats (TRs) occur in a substantial fraction of protein sequences and exhibit distinct structural, functional, and evolutionary properties compared with aperiodic regions that form globular folds. Although TRs often originate as highly periodic sequences, they can progressively diverge through mutation while retaining residual periodicity, complicating their reliable detection. Numerous computational tools have been developed for TR identification, but benchmarking studies show that no single method performs optimally across all repeat classes, as performance varies depending on repeat type and length. To address these limitations, we developed Meta Repeat Finder, a comprehensive framework that integrates multiple TR detection methods, each optimized for specific repeat length ranges, into a unified pipeline. Meta Repeat Finder applies stringent validation criteria to the combined outputs of multiple detectors to identify structurally and functionally relevant TR regions. The pipeline supports standardized integration of new TR detection tools under defined criteria, ensuring consistent scoring, terminal repeat trimming followed by TR re-assessment, and generation of a non-redundant TR set. Results are presented through a web interface at https://bioinfo.crbm.cnrs.fr/mrf offering graphical visualization and downloadable tabular data. A command-line implementation is also available for large-scale analyses.