Chromosight

Chromosight detects higher-order organizational features in chromosomes by identifying patterns such as loops and self-interacting domains in Hi-C contact maps to support analysis of spatial genome organization.

Key Features:

• Input data: Operates on Hi-C contact maps.
• Pattern detection: Detects distinct patterns such as self-interacting domains and chromatin loops.
• Algorithmic approach: Implements a pattern-detection algorithm inspired by computer vision principles.
• Sensitivity: Exhibits enhanced sensitivity compared to existing methods, particularly on synthetic simulated data.
• Performance: Operates efficiently, providing faster detection while maintaining accuracy.
• Organism scope: Applicable to genomes from bacteria, viruses, yeasts, and mammals.
• Training requirement: Does not require prior training datasets and functions with default parameters.

Scientific Applications:

• Loop and domain identification: Identification of chromatin loops and self-interacting domains in Hi-C data.
• Comparative genomics: Comparative analysis of chromosomal architecture across bacteria, viruses, yeasts, and mammals.
• Spatial genome organization: Study of higher-order spatial genome organization and its links to biological function.
• Benchmarking: Evaluation and benchmarking of detection sensitivity using synthetic simulated Hi-C datasets.

Methodology:

Applies a computer vision-inspired pattern-detection algorithm to Hi-C contact maps to detect loops and self-interacting domains; operates without prior training using default parameters and has been evaluated on synthetic simulated data.