Felix Becker

Extended Tiberius models outscore Helixer and ANNEVO on all 33 genomes from the five clades all three tools cover, and approach BRAKER3 accuracy in four clades without extrinsic evidence.

Vipsania is an unsupervised deep learning model for gene finding in eukaryotic genomes, achieving higher accuracy than supervised methods across most clades.

Tiberius is a deep learning gene predictor that integrates HMMs with CNNs and LSTMs, achieving high accuracy in mammalian genomes.

VARUS samples RNA reads from the Sequence Read Archive to enable automated genome annotation with high sensitivity and specificity.