4DNvestigator: time series genomic data analysis toolbox

Abstract

Data on genome organization and output over time, or the 4D Nucleome (4DN), require synthesis for meaningful interpretation. Development of tools for the efficient integration of these data is needed, especially for the time dimension. We present the '4DNvestigator', a user-friendly network-based toolbox for the analysis of time series genome-wide genome structure (Hi-C) and gene expression (RNA-seq) data. Additionally, we provide methods to quantify network entropy, tensor entropy, and statistically significant changes in time series Hi-C data at different genomic scales.

Keywords: 4DN; centrality; entropy; networks; time series.

Figure 1.

The 4D Nucleome. (a) representative time series Hi-C and RNA-seq data correspond to genome structure and function, respectively. (b) genome structure and function are intimately related. the 4DNvestigator integrates and visualizes time series data to study their dynamical relationship

Figure 2.

Overview of the 4DNvestigator data processing pipeline. within this diagram, 4DN refers to the 4DNvestigator

Figure 3.

4DN feature analyzer. (A) 4DN data is input to the 4DN feature analyzer. Top: Structure data (Hi-C) is described using one-dimensional features for compatibility with function data (RNA-seq). Bottom: Multiple structural features and function data are integrated into the structure-function feature matrix. (B) The 4DN feature analyzer can use structure and function data directly to visualize a system’s dynamics using the 4DN phase plane [3,15]. Structure defines the $x$-axis (left: eigenvector centrality, right: PC1) and function defines the $y$-axis (left: $\mathrm{l}\mathrm{o}{\mathrm{g}}_2$(RPKM), right: $\mathrm{l}\mathrm{o}{\mathrm{g}}_2$(TPM)), and points show structure-function coordinates through time. Left: Maternal and paternal alleles of nine cell cycle genes through G1, S, and G2/M phases of the cell cycle (adapted from [15]). Right: Top ten genomic regions (100 kb) with the largest changes in structure and function during cellular reprogramming [7]. (C) Multiple dimension reduction techniques can be used to visualize the 4DN feature analyzer’s structure-function feature matrix (from left to right: LE, UMAP, and t-SNE). Top: 100 kb regions of Chromosome 4 across six time points during cellular differentiation [28]. Bottom: 100 kb regions of Chromosome 11 across three time points during cellular reprogramming [7]. (D) Example output of the 4DN feature analyzer. The output includes genes contained in loci with the largest changes, and links to their NCBI and GeneCards database entries [8,9]