NetDes 1.0.4

Network modeling based on dynamic equation simulation (NetDes)

NetDes

Network inference and optimization using Dynamical equation simulations

NetDes is a computational method for optimizing gene regulatory networks (GRNs) of core transcription factor (TF) based on gene expression time trajectories. NetDes was specifically designed for analyzing time-series scRNA-seq data. The NetDes pipeline contains the following steps.

• (1) NetDes calculates smoothed gene expression trajectories along inferred pseudotime.

• (2) Genes are then clustered according to the trajectories.

• (3) Core TFs are inferred for each gene cluster using gene set analysis and TF-target gene relationship.

• (4) An initial GRN of core TFs is constructed according to TF-target gene relationship, where target genes also belong to the core TFs.

• (5) Network optimization that refines the GRN and constructs a dynamical model using ordinary differential equations (ODEs).

• (6) Dynamical systems modeling, such as gene perturbation simulations, signal driving simulations, network coarse-graining.

Installation

Python version 3.9 or greater is required.

Install from PyPi (recommended)

To install the most recent release, run

pip install NetDes

Install with github

• Git clone the NetDes repository, cd to the NetDes directory, and run

pip install .

Tutorials

[Processing gene expression time trajectories using scRNA-seq data]: This script illustrates the steps to process input scRNA-seq data, obtain smoothed gene expression time trajectories, and cluster genes based on the trajectories. (Steps 1 and 2)

[Inferring core TFs]: This script shows the inference of core transcription factors using Fisher's exact test. (Step 3)

[Initial GRN construction]: This script illustrates how to build an initial GRN using TF-target gene databases like Rcistarge, TRRUST, and NetAct. (Step 4)

[GRN optimization and simulation]: This tutorial page shows the NetDes' usage for network optimization and simulation. The user input includes pseudotime, smoothed gene expression trajectories, and an initial GRN.

[Benchmarking]: This script explains the steps for GRN evaluation by comparing RACIPE simulations with the scRNA-seq data. (Step 5)

[Network coarse-graining]: This script shows the process of coarse-graining the optimized GRN into a small gene circuit using SacroGraci. (Step 6)