Differential analysis of image-based chromatin tracing data with Dory

The spatial organization of the genome plays a vital role in maintaining the cell type-specific chromatin structures and regulating gene expression. The three-dimensional (3D) genome structures can be measured by sequencing-based techniques such as Hi-C usually on the ensemble level or by imaging-based techniques such as chromatin tracing on the single-molecule level. Chromatin tracing is a multiplexed DNA fluorescence in situ hybridization (FISH) technique that can directly trace the 3D positions of genomic loci along individual chromosomes in single cells. However, few computational tools have been developed for statistical differential analysis of chromatin tracing data. Major challenges include the high dimensional nature of the data, high variability of chromatin traces and missing values in the 3D coordinates of some loci. Here, we present Dory, a statistical method for identifying differential patterns from two groups of chromatin traces without imputation of missing data. Dory measures the spatial distance between each pair of regions in each chromatin trace and employs multi-level statistical tests to quantify the difference between the two groups of chromatin traces. The output of Dory includes the differential score for each region pair comparing two biological conditions and significantly differential region pairs. We applied Dory to chromatin tracing data from six cell types in mouse fetal liver and from a mouse lung cancer progression model. We found that the changes in chromatin structure are associated with A/B compartment alterations and promoter-enhancer interaction alterations that cause differential gene expression. Dory is a user-friendly computational tool that can be applied to multiple types of imaging-based 3D genome data analysis for studying genome functions and gene regulation.