HINT

HINT identifies active transcription factor (TF) binding sites by integrating DNase I hypersensitivity (DHS) and histone modification occupancy using hidden Markov models (HMMs) to enable cell-specific prediction of regulatory elements.

Key Features:

• Integration of Chromatin Data: Combines DNase I hypersensitivity (DHS) and histone modification occupancy using hidden Markov models (HMMs) to detect open chromatin regions and active TF binding sites.
• Model Framework: Implements genomic signal processing, model training, and genome-wide application within an HMM framework for regulatory element identification.
• Performance Metrics: Optimizes sensitivity and specificity and reports superior area under the curve (AUC) statistics in comparative analyses.
• No Additional Training Required: Generates predictions for new cell types without requiring additional training data or sequence information.
• Flexibility for Downstream Analysis: Outputs active binding sites suitable for downstream analyses such as de novo motif finding.
• Regulatory Genomics Toolbox: Provided as part of the Regulatory Genomics Toolbox.

Scientific Applications:

• Regulatory network elucidation: Mapping active TF binding sites to reconstruct transcriptional regulatory networks.
• Development and disease research: Investigating regulatory mechanisms underlying cell development and disease onset using cell-specific TF binding predictions.

Methodology:

Utilizes hidden Markov models (HMMs) to integrate DNase I hypersensitivity (DHS) and histone modification data; performs genomic signal processing, model training, and genome-wide application; and evaluates predictions using sensitivity, specificity, and AUC statistics.