Single-cell network biology characterizes cell type gene regulation for drug repurposing and phenotype prediction in Alzheimer’s disease

Summary

Dysregulation of gene expression in Alzheimer’s disease (AD) remains elusive, especially at the cell type level. Gene regulatory network, a key molecular mechanism linking transcription factors (TFs) and regulatory elements to govern target gene expression, can change across cell types in the human brain and thus serve as a model for studying gene dysregulation in AD. However, it is still challenging to understand how cell type networks work abnormally under AD. To address this, we integrated single-cell multi-omics data and predicted the gene regulatory networks in AD and control for four major cell types, excitatory and inhibitory neurons, microglia and oligodendrocytes. Importantly, we applied network biology approaches to analyze the changes of network characteristics across these cell types, and between AD and control. For instance, many hub TFs target different genes between AD and control (rewiring). Also, these networks show strong hierarchical structures in which top TFs (master regulators) are largely common across cell types, whereas different TFs operate at the middle levels in some cell types (e.g., microglia). The regulatory logics of enriched network motifs (e.g., feed-forward loops) further uncover cell type-specific TF-TF cooperativities in gene regulation. The cell type networks are highly modular and several network modules with cell-type-specific expression changes in AD pathology are enriched with AD-risk genes and putative targets of approved and pending AD drugs, suggesting possible cell-type genomic medicine in AD. Finally, using the cell type gene regulatory networks, we developed machine learning models to classify and prioritize additional AD genes. We found that top prioritized genes predict clinical phenotypes (e.g., cognitive impairment) with reasonable accuracy. Overall, this single-cell network biology analysis provides a comprehensive map linking genes, regulatory networks, cell types and drug targets and reveals dysregulated cell type gene dysregulatory mechanisms in AD.

Flow chart

System requirements

The analysis is based on R version 4.0.3. For the gene regulatory network, a Linux system with 32 GB RAM and 32GB storage would be enough. For network analysis, a standard computer should be enough.

Software installation

Packages needed for the whole analysis.

igraph v1.2.6 (R package, https://igraph.org/r/)
mfinder v1.21 (motif finding tool from https://www.weizmann.ac.il/mcb/UriAlon/download/network-motif-software)
Loregic v1 (R package for regulatory logics, https://github.com/gersteinlab/Loregic.git)
WGCNA 1.70-3 (R package for network module detection, https://horvath.genetics.ucla.edu/html/CoexpressionNetwork/Rpackages/WGCNA/)
randomForest 4.6-14 (R package for binary classification)
other supporting packages should install automatically. Check load_libraries.R file in the scripts directory

Download code

The code has been tested on R version 4.0.3 on Linux and Mac OS.

git clone https://github.com/daifengwanglab/scNET
cd scNET

Download data

Download brain cell type gene regulatory networks for four cell types from Zenodo (DOI:10.5281/zenodo.5829585) and store them in the data directory. These networks link TFs to target genes based on open chromatin data, chromatin interaction maps, and gene expression correlations in control and AD cells. scGRNom pipeline was used to predict the networks.

cd ADnets/
mkdir results

Run the following lines of code in the R console. The outputs will be stored in the results directory

1. Get gene centrality

source('../scripts/get_centrality.R')
source('../scripts/get_h_metric.R')

2. Get network motifs

This analysis is run outside of R using the mfinder tool. Please refer to mfinder manual on the link provided above. Note that mfinder uses numeric gene IDs. The script convert_symbols_to_entrezID.R will assign numeric IDs to gene symbols and generate a *.node_symbol-integar.key.txt file for this mapping to be used later.

Rscript ../scripts/convert_symbols_to_entrezID.R data/MIT.AD.Mic.grn.txt Mic.AD
/path/to/mfinder Mic.Ctrl.TF-TF.entrez.txt -s 3 -r 1000 -f Mic.AD -ospmem 38
mv Mic.AD_MEMBERS.txt results/
mv Mic.AD_MEMBERS.txt.node_symbol-integar.key.txt results/
rm Mic.AD.TF-TF.entrez.txt

3. Get regulatory logics

An example of how to get regulatory logics for a single cell type GRN using Loregic. Loregic requires a binarized gene expression matrix. Please follow Loregic's documentation to create such a matrix. A toy matrix called Mic.AD.binned_expr for the purposes of this demo is available in the data directory.

Rscript run_loregic_FFL.R results/Mic.AD_MEMBERS.txt results/Mic.AD_MEMBERS.txt.node_symbol-integar.key.txt Mic AD data/Mic.AD.binned_expr 99

4. Get co-regulatory network modules

source('../scripts/get_module_coregnet.R')

5. Prioritize network genes using random forest classifier

source('../scripts/network_based_classifier.R')