Chromosomes of all species studied so far display a variety of higher-order organisational features, such as self-interacting domains or loops. These structures, which are often associated to biological functions, form distinct, visible patterns on genome-wide contact maps generated by chromosome conformation capture approaches such as Hi-C. Here we present Chromosight, an algorithm inspired from computer vision that can detect patterns in contact maps. Chromosight has greater sensitivity than existing methods on synthetic simulated data, while being faster and applicable to any type of genomes, including bacteria, viruses, yeasts and mammals. Our method does not require any prior training dataset and works well with default parameters on data generated with various protocols.

Chromosomes of all species studied so far display a variety of higher-order organizational features such as domains, loops, or compartments. Many of these structures have been characterized from the genome-wide contact maps generated by chromosome conformation capture approaches (Hi-C, ChIA-PET,...). Indeed, DNA 3D structures translate as distinct patterns visible on these maps. We developed Chromosight, an algorithm based on computer vision approaches that automatically detect and quantify any type of pattern in contact data. Chromosight detects 3 times as many patterns as existing programs, while being faster and fit to any genome, including small, compact ones. Chromosight is user-friendly and can be extended to user-provided patterns. We validated the program by applying it to a variety of chromosomal structures found in mammals. Code and documentation: https://github.com/koszullab/chromosight