WORMHOLE

WORMHOLE predicts least diverged orthologs (LDOs) by integrating outputs from multiple ortholog prediction algorithms using machine learning to improve ortholog identification across species for comparative biology and functional inference.

Key Features:

• Integration of Multiple Algorithms: Combines predictions from 17 distinct ortholog prediction algorithms to aggregate evidence for LDO identification.
• Machine Learning Application: Applies a machine learning framework to synthesize diverse predictive strategies and increase confidence in LDO calls.
• Cross-Species Analysis: Analyzes ortholog relationships among six eukaryotic species — humans, mice, zebrafish, fruit flies, nematodes, and budding yeast — to improve LDO prediction between distantly related taxa while prioritizing low evolutionary distance and high functional similarity.
• Validation and Performance: Performs extensive validation including cross-validated predictions of PANTHER LDOs and evaluation of evolutionary divergence and functional similarity to assess prediction robustness.

Scientific Applications:

• Comparative biology: Identifies orthologous genes with minimal evolutionary divergence to support comparative genomic analyses.
• Human disease models: Facilitates selection of functionally conserved orthologs for modeling human disease in model organisms such as mice, zebrafish, fruit flies, nematodes, and budding yeast.
• Cross-species functional inference: Enables identification of functionally related gene pairs across species to inform studies of conserved genetic functions and disease mechanisms.

Methodology:

Integrates outputs from 17 ortholog prediction algorithms into a machine learning framework, analyzes combinations of query and target species, and uses cross-validation against PANTHER LDOs while evaluating evolutionary divergence and functional similarity.