4DNvestigator Toolbox

Overview

The 4DNvestigator is a MATLAB toolbox that analyzes time-series genome-wide chromosome conformation capture (Hi-C) and gene expression (RNA-seq) data.

Paper: in preparation

Availability: https://github.com/lindsly/4DNvestigator

Data Availability: 4DNvestigator Data

Installation

Install the 4DNvestigator through one of the following methods. Both methods will create a directory named 4DNvestigator which will contain all the necessary files for execution.

• Download the toolbox as a .zip folder and extract the contents into your MATLAB directory of choice.
• Clone this Git repository into your MATLAB directory of choice using MATLAB's built-in Source Control

After the 4DNvestigator directory (and its sub-directories) are added, all example data must be downloaded to the folder "data\exampleData" in order to run all of the functions contained in the example script "ExampleScript.m"

Hi-C and RNA-seq data types

The 4DNvestigator accepts the following input file formats:

Core Functions

• 4DN Feature Analyzer: This measures the amount of change in both genome structure and function for specified genomic regions by mapping all time points to a consistent low dimensional embedding, and quantifying the variance of each loci within this space over time. Method specifics can be found in: "Genome Architecture Mediates Transcriptional Control of Human Myogenic Reprogramming"
• von Neuman Entropy: Measures the entropy ("uncertainty") of a multivariate system. Uncertainty is related to stemness. Here, we use this measure to determine stemness of Hi-C samples.
• Larntz-Perlman: Method for testing the equality of correlation matrices. This is applied to Hi-C correlation matrices to determine the significance of matrix differences.
• Network Analysis: Measures the overlapping degree and multiplex participation coefficient to evaluate the heterogeneity of nodal degrees in dynamic biological networks.

Other Functions

• Chromatin partitioning: Partitioning of the genome into two distinct groups based on either the Fiedler vector or principal component 1. This partitioning corresponds to euchromatin and heterochromatin, or A/B compartments.
• Differential expression: Differential expression measures the significance of RNA-seq expression differences between samples.
• A/B switching: This function determines which genomic regions change their chromatin structure from compartment "A" to compartment "B"

Examples

• See "ExampleScript.m" to run each of the core functions of the 4DNvestigator with default parameters using provided example data.
• GettingStarted_4DNvestigator provides further details for each core function and their respective output.

Dependencies

• The 4DNvestigator requires that java and python are installed to extract data from .hic files.
  • We recommend that the "requests" library is installed for python, and that it is added to the MATLAB path with "setenv"
• The 4DNvestigator uses juicertools to read data from .hic files
• The 4DNvestigator requires the following MATLAB Toolboxes for proper functionality: Statistics and Machine Learning, Computer Vision, Bioinformatics, and Image Processing

System Requirements

We recommend the following system properties for all 4DNvestigator functionalities to run properly:

• Windows 10
• At least 12 Gb RAM
  • This is highly dependent on the size of the data being analyzed (data used in "ExampleScript.m" peaks at 9.5 Gb RAM usage)
• A current version of MATLAB (2019 or later)

4DNvestigator was written and tested on a Windows 10 machine with an Intel Core i7-8700 CPU and 32 Gb RAM using MATLAB R2019b.