EagleC

EagleC detects structural variations in human genomes from Hi-C contact map data using deep-learning and ensemble-learning to produce high-resolution predictions across a full spectrum of genomic alterations.

Key Features:

• Full-Range Detection: Detects a comprehensive range of SVs including interchromosomal translocations and intrachromosomal variants across scales and SVs that can be missed by whole-genome sequencing and nanopore sequencing.
• High Resolution: Provides high-resolution identification of subtle and complex structural variants.
• Multi-platform Support: Processes chromatin interaction data from Hi-C, HiChIP, ChIA-PET, and capture Hi-C for SV detection.
• Single-Cell Application: Applies to single-cell Hi-C data to assess SV heterogeneity within primary tumors.
• Deep-learning and Ensemble Learning: Employs advanced deep-learning models combined with ensemble-learning techniques for robust SV prediction.

Scientific Applications:

• Cancer Genomics: Applied to over 100 cancer cell lines and primary tumors to identify high-confidence structural variants implicated in cancer.
• Fusion Gene Discovery: Recovers fusion genes and other structural variants that contribute to tumor biology.
• Tumor Heterogeneity and Evolution: Enables exploration of SV heterogeneity and tumor evolution at single-cell resolution.
• Complementing Sequencing: Complements whole-genome sequencing and nanopore sequencing by detecting SVs those methods may miss.

Methodology:

Processes Hi-C contact map data with deep-learning models and aggregates outputs using ensemble-learning techniques to predict structural variations.