3D-GNOME 2.0: a three-dimensional genome modeling engine for predicting structural variation-driven alterations of chromatin spatial structure in the human genome

Abstract

Structural variants (SVs) that alter DNA sequence emerge as a driving force involved in the reorganisation of DNA spatial folding, thus affecting gene transcription. In this work, we describe an improved version of our integrated web service for structural modeling of three-dimensional genome (3D-GNOME), which now incorporates all types of SVs to model changes to the reference 3D conformation of chromatin. In 3D-GNOME 2.0, the default reference 3D genome structure is generated using ChIA-PET data from the GM12878 cell line and SVs data are sourced from the population-scale catalogue of SVs identified by the 1000 Genomes Consortium. However, users may also submit their own structural data to set a customized reference genome structure, and/or a custom input list of SVs. 3D-GNOME 2.0 provides novel tools to inspect, visualize and compare 3D models for regions that differ in terms of their linear genomic sequence. Contact diagrams are displayed to compare the reference 3D structure with the one altered by SVs. In our opinion, 3D-GNOME 2.0 is a unique online tool for modeling and analyzing conformational changes to the human genome induced by SVs across populations. It can be freely accessed at https://3dgnome.cent.uw.edu.pl/.

Figure 1.

A schematic representation of the workflow of 3D-GNOME 2.0.

Figure 2.

Schemes representing the behaviour of the computational algorithm implemented in 3D-GNOME 2.0 for prediction of changes in 3D chromatin contacts induced by SVs. (A) CTCF ChIA-PET contact diagrams for exemplary region chr1:47656996-48192898 containing TAL1 locus for the reference genome (GM12878) and upon introduction of SVs. Alteration of CTCF-mediated contact patterns upon addition of DUP, DEL, INV or INS to the genomic sequence is shown. SVs are marked as blue rectangles. CTCF anchors and their directionality are represented as white circles and arrows, correspondingly. Red arrows represent CTCF anchors’ alterations induced by SVs. (B) 3D models of CTCF mediated chromatin structures corresponding to genomic regions shown in (A). Loops are coloured as genomic regions represented below CTCF contact diagrams depicted in (A).

Figure 3.

Output of 3D-GNOME 2.0. for an exemplary region (chr14:35138000-36160000) affected by two deletions: (chr14:35401403-35401724 and chr14:35605439-35615196). (A) Screenshot of the result page with diagrams of chromatin contacts mediated by CTCF (green) and RNAPII (purple) for both the reference genome (based on GM12878 data) and the variant genome (based on SVs from HG00099); the deletion chr14:35605439-35615196 (right DEL) disrupts CTCF and RNAPII interactions. (B) Clusters length distribution for CTCF and RNAPII protein factors for reference genome (top) and HG00099 (bottom). (C, D) Representation of 3D models in the 3D viewer, proposed using only CTCF interactions (panel C) or both CTCF and RNAPII data (panel D) from the reference genome (left) and that affected by SVs (right).