PROFdisis

PROFdisis predicts protein–DNA interaction sites directly from amino acid sequences to identify residues that contact DNA without requiring structural or experimental data.

Key Features:

• Sequence-based prediction: Predicts DNA-contacting amino acid residues directly from primary sequence information.
• Biophysical feature modeling: Leverages distinct biophysical characteristics inherent to DNA-binding residues to inform predictions.
• Structure-independent: Operates without the need for 3D structural data or additional experimental input.
• Cross-validation framework: Employs a robust cross-validation scheme to assess predictive performance.
• Training dataset: Utilizes the largest non-redundant dataset of high-resolution protein–DNA complexes available for model development and evaluation.
• Performance validation: Achieves validation with 89% of its predictions confirmed against experimental data.

Scientific Applications:

• Proteome-scale interaction mapping: Enables large-scale identification of protein–DNA interaction sites across proteomes using sequence data alone.
• Genomics: Supports studies of DNA-binding proteins relevant to genome regulation and annotation.
• Proteomics: Assists in annotation of DNA-binding residues within protein datasets for functional proteomic analyses.
• Systems biology: Facilitates integration of predicted protein–DNA interactions into network and systems-level models.
• Functional studies: Informs research into gene regulation mechanisms and DNA repair processes by identifying candidate contact residues.

Methodology:

Sequence-based prediction using biophysical characteristics of DNA-binding residues, evaluated with a robust cross-validation framework on a large non-redundant set of high-resolution protein–DNA complexes, with 89% of predictions confirmed against experimental data.