ensECBS

ensECBS enhances chemical similarity searching by integrating machine learning models with evolutionary relationships of target genes to refine identification of small molecules with functional activity related to target binding.

Key Features:

• Machine Learning Integration: Employs multiple classification similarity-learning models that utilize various levels of evolutionary information about target genes.
• Evolutionary Relationship Utilization: Encodes evolutionary data of target genes into chemical representations through their binding targets to expand chemical-target interaction datasets.
• Probability-Based Similarity Scoring: Quantifies chemical similarity as the probability that two chemicals will bind to identical targets.

Scientific Applications:

• Enhanced Chemical Discovery: Identifies hidden chemical relationships and novel compounds with potential biological activity by leveraging evolutionarily conserved target-binding information.
• Improved Model Performance: Boosts similarity-model performance by integrating heterogeneous multiple target-binding data into paired formats for classification.

Methodology:

Processes comprehensive chemical-target interaction datasets formatted as paired data and analyzes them using classification similarity-learning models that incorporate different levels of target evolutionary information, producing probability-based similarity scores for shared target binding.