DDNA

DDNA predicts protein–DNA binding affinity and identifies DNA-binding proteins by combining structural alignment with a DFIRE-based statistical energy function incorporating a volume-fraction correction.

Key Features:

• Volume-fraction correction: Integrates a volume-fraction correction into the statistical energy function to improve binding affinity estimates.
• DFIRE-based energy function: Uses a distance-scaled, finite, ideal-gas reference (DFIRE) statistical energy model to evaluate protein–DNA interaction energies.
• Structural alignment (TM-align): Employs TM-align for structural alignment to ensure comparisons among structurally similar proteins.
• Template-based prediction: Addresses template-based prediction of DNA-binding proteins by combining structural similarity and energetic evaluation.
• Predictive performance: Reports a Matthews correlation coefficient (MCC) of 0.76 (previous methods up to 0.69) with accuracy 98%, precision 93%, and sensitivity 64%.

Scientific Applications:

• Structural genomics screening: Applied to 2,235 structural genomics targets to identify candidate DNA-binding proteins, yielding 37 potential binders.
• Functional annotation: Supports functional characterization of proteins with unknown roles by predicting DNA-binding capability, with 27 of 37 predicted candidates (73%) confirmed as putative DNA binders in the applied dataset.

Methodology:

DDNA performs structural alignment with TM-align and evaluates protein–DNA interactions using a DFIRE-based statistical energy function that incorporates a volume-fraction correction.