PanTEon: a cross-kingdom framework to guide the design of transposable element classifiers

PanTEon reveals that machine learning-based transposable element classifiers show markedly variable performance across TE superfamilies and eukaryotic lineages.

• Why it matters: Inconsistent TE annotation across species hampers comparative genomics and progress in TE biology, highlighting the need for standardized, reproducible classification tools.
• What they did: A cross-kingdom deep learning framework, PanTEon, was developed with a broad TE database and benchmarking platform, evaluating nine machine learning methods across animals, plants, and fungi.
• The result: Ensemble and phylum-specific models improve predictive F1 scores, but cross-species generalization remains difficult, providing a scalable foundation for future TE classification and community efforts.