Statistically rigorous and computationally efficient chromatin stripe detection with Quagga.

Chromatin stripes are architectural chromatin features in which a singular loop anchor interacts with a contiguous region of DNA so, at the bulk sequencing level, it appears as a long stripe on chromatin contact matrices. Stripes are thought to play an important role in gene regulation and have been implicated in regulating a cell's lineage determination. Therefore, integrated analysis of stripes with genomic and epigenomic features at a genome-wide scale shows vast potential in understanding the cooperation between regulatory elements in 3D nucleome. To this end, we present Quagga, a computational tool for detection and statistical verification of genomic architectural stripes from Hi-C or Micro-C chromatin contact maps, that relies on robust image processing techniques and rigorous statistical tests for enrichment. Quagga outperforms other stripe detection methods in accuracy and is highly versatile, working with Hi-C, Micro-C, and other chromatin conformation capture data. By reporting on all tools' performance in classifying CTCF-cohesin anchored stripes, enhancer-promoter interacting stripes, and indeterminate stripes, we also demonstrate a thorough, integrated analysis to determine the output stripes' quality. Our work provides a flexible and convenient tool to help scientists explore the relationships between chromatin architectural stripes and important biological questions.