ChINN

ChINN predicts chromatin interactions from DNA sequences to infer CTCF-, RNA polymerase II–associated, and Hi-C–detectable contacts among open chromatin regions.

Key Features:

• Sequence-Based Prediction: Uses DNA sequence input to predict chromatin interactions without requiring genome-wide interaction assays.
• Comprehensive Interaction Types: Predicts interactions associated with CTCF and RNA polymerase II and targets interactions detectable by Hi-C.
• Performance Across Samples: Demonstrates robust predictive performance across different samples by capturing diverse sequence features associated with interaction sites.
• Application in Heterogeneous Data Sets: Applied to datasets including six chronic lymphocytic leukemia (CLL) patient samples and an additional cohort of 84 CLL open chromatin samples to identify heterogeneity in chromatin interactions.

Scientific Applications:

• Gene Regulation Studies: Infers regulatory architecture and gene regulatory mechanisms by predicting physical contacts between regulatory elements from sequence data.
• Disease Genomics (CLL): Characterizes chromatin interaction heterogeneity in chronic lymphocytic leukemia (CLL) using open chromatin samples.

Methodology:

Uses a neural network machine learning approach to analyze DNA sequences and integrate sequence features for prediction of chromatin interactions.