PiDNA

PiDNA predicts protein-DNA interactions using an atomic-level knowledge-based scoring function applied to in silico mutated complex structures to infer transcription factor binding preferences and construct position weight matrices (PWMs).

Key Features:

• Atomic-Level Scoring Function: Employs an atomic-level knowledge-based scoring function applied to numerous in silico mutated protein-DNA complex structures to assess interaction affinities and specificities.
• Position Weight Matrix (PWM) Construction: Constructs PWMs from structural models exhibiting minimal energy changes to reflect protein-DNA interaction dynamics consistent with experimental data.
• Prediction Capabilities: Predicts relative preferences for all DNA sequences with limited mutations from native sequences co-crystallized in models in a single run and supports broader exploration involving unlimited mutations via additional input models or files.
• Outputs: Produces a ranked list of mutated sequences, the PWM derived from favorable mutated structures, and specific mutated protein-DNA complex structure models as output.
• Validation and Accuracy: Validated against in vitro protein-binding microarray (PBM) data, demonstrating detection of high-specificity binding sites and selection of experimentally validated sites from large sequence pools.
• Experimental Design and Protein Design Support: Provides sequence-specificity predictions and mutated structure models to inform biological experiment design and downstream protein design applications.
• Integration with ChIP Data: Can be integrated with chromatin immunoprecipitation (ChIP) data to refine large-scale inference of in vivo protein-DNA interactions.

Scientific Applications:

• Transcription Factor Binding Site Prediction: Ranks and predicts transcription factor binding preferences from structural complexes and mutated sequence variants.
• Experimental Design and Hypothesis Testing: Supplies sequence-specificity predictions and mutated complex models to guide design and interpretation of molecular biology experiments.
• Protein Engineering and Design: Provides mutated protein-DNA complex models and specificity data to support protein engineering and design efforts.
• Integration with Genomic Data: Combines PWMs and structural predictions with chromatin immunoprecipitation data to improve inference of in vivo interactions.

Methodology:

Constructs PWMs from structural models by applying an atomic-level knowledge-based scoring function to in silico mutated protein-DNA complex structures.