ProteinFuncPred

ProteinFuncPred predicts protein functions from domain architecture by modeling domain recurrence and domain order to probabilistically associate proteins with Gene Ontology terms.

Key Features:

• Domain Recurrence and Order: Integrates domain recurrence and the sequential arrangement of domains within proteins to inform function prediction.
• DRDO (Domain Recurrence and Domain Order): Computes posterior probabilities of Gene Ontology (GO) terms by leveraging information on domain recurrence and domain order.
• DRDO-NB (Naïve Bayes Approach): Applies a Naïve Bayes model using domain architecture features to predict protein functions.
• Improved Accuracy: Large-scale benchmark comparisons demonstrate significant improvement in protein function inference accuracy relative to traditional approaches.

Scientific Applications:

• Protein Function Annotation: Annotates proteins from genomic and proteomic datasets by predicting functions based on domain architecture.
• Gene Ontology Term Prediction: Provides probabilistic associations between proteins and specific Gene Ontology terms to enhance functional characterization of proteomic data.

Methodology:

Two complementary computational approaches are used: DRDO computes posterior probabilities of GO-term associations from observed patterns of domain recurrence and order, and DRDO-NB applies a Naïve Bayes model using domain architecture features to predict protein functions.