gitter-lab / SINGE-supplemental

Supplemental scripts and analyses for the SINGE manuscript

SINGE supplemental information

This repository contains supplemental scripts and analyses for the SINGE manuscript:

Network inference with Granger causality ensembles on single-cell transcriptomics.
Atul Deshpande, Li-Fang Chu, Ron Stewart, Anthony Gitter.
Cell Reports, 38:6, 2022.

See the main SINGE repository for the software.

Contents

• Commands for running network inference algorithms
• dyngen: script and results for simulated gene regulatory networks
• escape: scripts to prepare the ESCAPE gold standard
• files: supplementary files associated with the manuscript
• scripts: scripts to execute the methods used in the manuscript and plot the figures in the manuscript

Usage documentation of SINGE and other network inference methods

For network inference from the different methods, we use the following commands. See the scripts directory for related scripts.

SINGE

For all datasets in the paper, we first run all GLG tests for each target using

bash SINGE.sh PATH_TO_RUNTIME GLG DATA_FILE GENELIST_FILE OUTPUT_FOLDER default_hyperparameters.txt HYPENUM

Inputs

• DATA_FILE is a version -v7.3 mat-file containing an expression matrix X of dimensions GxC, where G is the number of genes and C is the number of cells, and the vector ptime of dimensions 1xC, containing the pseudotimes assigned to each cell.
• GENELIST_FILE contains a cell-array containing the names of the G genes corresponding to the entries in X.
• default_hyperparameters.txt is available in the SINGE repository https://github.com/gitter-lab/SINGE.
• HYPENUM is the hyperparameter number, ranging from 1 to 200 in this case.

Outputs

Each GLG Test gives an output adjacency matrix A(HYPENUM) representing a partially inferred network saved in the OUTPUT_FOLDER.

This is followed by an aggregation of all GLG Test results using

bash SINGE.sh PATH_TO_RUNTIME Aggregate DATA_FILE GENELIST_FILE OUTPUT_FOLDER

Inputs

• OUTPUT_FOLDER and its contents.

Outputs (saved in the OUTPUT_FOLDER)

• SINGE_Ranked_Edge_List.txt : File with list of ranked edges according to their SINGE scores.
• SINGE_Gene_Influence.txt : File with list of genes ranked according to their SINGE influence.

SINCERITIES

The SINCERITIES MATLAB version 1.0 package is available at http://www.cabsel.ethz.ch/tools/sincerities.html, and we directly use its graphical user interface (MAIN.m) with default settings for each dataset.

SCODE

We downloaded SCODE from the GitHub repository https://github.com/hmatsu1226/SCODE (git commit 28acad67893c0fba7eeee670c339809d45ae6377) and used the command and settings listed below

ruby run_R.rb <Input_file1> <Input_file2> <Output_dir> <G> <D> <C> <I> <R>

Inputs

• Input_file1 : G x C matrix of expression data
• Input_file2 : Time point data (e.g. pseudo-time data)
• Output_dir : Folder to store temporary output files
• G : number of genes (obtain from dataset)
• D : dimensionality of vector z (D=4 for ESC to Endoderm Differentiation dataset, D=20 for Retinoic Acid-driven Differentiation dataset)
• C : number of cells in the dataset (obtain from dataset)
• I : number of iterations (I=100)
• R : number of trials over which SCODE output is averaged (R=1000).

Output

• meanA.txt : Represents the adjacency matrix obtained from the average output of R SCODE trials

Jump3

We parallelize the Jump3 implementation to speed up the processing time using high-throughput computing. An equivalent version of the Jump3 code we used can be obtained from https://github.com/vahuynh/Jump3 (git commit 03a7e86d82f2383c56fd11c658dfce574fbf1a1a). Because Jump3 did not terminate in a reasonable amount of time on the full Retinoic Acid-driven Differentiation dataset, we reduced the dataset by arbitrarily dropping cells with probability 0.5. We use only the ordering information in Jump3 with the assigned obsTimes increasing from 0 to C-1 from the earliest cell to the latest cell, where C is the number of cells in the dataset.

w = jump3(data,obsTimes,noiseVar,tfidx,K,ntrees)

We used noiseVar.obsnoise=0.1, K = 100, ntrees = 100, and replicate the default values from Jump3's example in its GitHub repository for the other variables.

GENIE3

We installed GENIE3 version 1.6.0 from the BioConductor page https://bioconductor.org/packages/release/bioc/html/GENIE3.html and used its default operating mode without any modifications.