ENTAIL

ENTAIL predicts fibril deposits associated with amyloidoses by using molecular descriptors and a Naive Bayes classifier to identify protein precursors involved in amyloid fibril formation.

Key Features:

• Molecular descriptors: Leverages over 4,000 molecular descriptors to characterize proteins for prediction.
• Classifier: Employs a Naive Bayes Classifier with Unbounded Support combined with a Gaussian Kernel Type.
• Predictive target: Identifies potential protein precursors that may contribute to amyloid fibril deposits.
• Dataset handling: Demonstrated efficacy on balanced datasets.
• Performance metrics: Reports accuracy 81.80%, sensitivity (SN) 100%, specificity (SP) 63.63%, and Matthews Correlation Coefficient (MCC) 0.683.

Scientific Applications:

• Identification of protein precursors: Detects proteins that may act as precursors in amyloidogenesis.
• Study of amyloid-associated diseases: Supports investigation of fibrillar deposits implicated in Alzheimer's disease, Creutzfeldt-Jakob disease, and type II diabetes.
• Insight generation: Provides predictive data that may inform understanding of pathological processes and future therapeutic strategies for amyloidoses.

Methodology:

Uses over 4,000 molecular descriptors and a Naive Bayes Classifier with Unbounded Support combined with a Gaussian Kernel Type, evaluated on balanced datasets with reported accuracy, sensitivity, specificity, and MCC.