📌 Please Note

A subset of the samples used in this paper overlap with samples used in Mathys et al (Nature, 2019). These samples were sequenced using V2 chemistry. Updated cell-level files (post-quality control) and individual-level metadata with this additional information are available for download from the Open Science Framework: cell-level data and individual-level metadata.

For supplemental analyses considering the chemistry differences, please refer to this Google Colab notebook: Supplemental Analyses Notebook.

This repository contains instructions for open data access and code to reproduce the analyses presented in

APOE4 impairs myelination via cholesterol dysregulation in oligodendrocytes

• Paper
• Nature Podcast
• Nature News
• News and Views

snRNAseq & lipidomic Data availability

• If you would like to process the raw counts matrix and associated metadata, the fully-processed, annotated, and qc-ed data, or the annotated pre-qc lipidomic data, these files can be downloaded from the AD Knowledge Portal under Synapse ID syn38120890. FASTQ files can also be found under the same Synapse ID.

other Data availability

• All other datasets needed for this analysis are available through the open science framework.
• each file is described in detail below

Reproduce analyses and plots

Follow these instructions to reproduce analyses and plots, as shown in the paper.

1. Download the repository by running:

git clone https://github.com/djunamay/apoe4myelin.git

2. Create the conda environments and install dependencies:

To create the first conda environment, run this in your terminal:

conda env create -f ./environment/apoe_env.yml
conda activate apoe4myelin_env
R

In your R console, run:

install.packages("BiocManager")
BiocManager::install(version=“3.14”)
BiocManager::install("ComplexHeatmap", version="3.14") # preferred package version = 2.10.0
BiocManager::install("BiocParallel", version="3.14") # preferred package version = 1.28.3
BiocManager::install("SingleCellExperiment", version="3.14") # preferred package version = 1.16.0
BiocManager::install("limma", version="3.14") # preferred package version = 3.50.3
BiocManager::install("GSVA", version="3.14") # preferred package version = 1.42.0
BiocManager::install("edgeR", version="3.14") # preferred package version = 3.36.0
BiocManager::install("fgsea", version="3.14") # preferred package version = 1.20.0

Create the second custom conda environment by running:

conda create -n actionet_legacy_env  -c conda-forge r-devtools

Install the ACTIONet package (legacy version) from github by running the following:

conda activate actionet_legacy_env
R

In your R console, run this:

library(devtools)
devtools::install_github("shmohammadi86/ACTIONet_legacy")
install.packages("harmony")
install.packages("R.utils")
install.packages("BiocManager")
BiocManager::install("BiocParallel")
install.packages("ggpubr")
BiocManager::install("scran")
BiocManager::install("scater")

3a. If you'd like to perform your own QC and celltype annotation from scratch

1. Download the raw counts matrix and associated metadata from Synapse here. Please note, a data-use agreement must be submitted to access these data. Follow instructions on Synapse here.

3b. If you'd like to recapitulate our QC and celltype annotation

1. create actionet_legacy_env conda environment as described in point 2
2. Follow instructions from point 3c.5 to access and download the /single_cell_data files marked by *. Download the raw_counts_data folder from Synapse, as described in point 3a.1.
3. Navigate to the scripts folder, Now run:

conda activate actionet_legacy_env
Rscript ../scripts/generate_raw_sce_object.r 
Rscript ../scripts/qc_and_annotation.r 

3c. If you'd like to recapitulate any of the analyses presented in the paper

1. create apoe4myelin_env conda environment as described in point 2
2. follow instructions here to download the nebula package into the apoe4myelin_env environment
3. follow instructions here to download the immunogenomics/presto package into the apoe4myelin_env environment
4. Create the following /plots directory within this git repo

apoe4myelin
└───plots
    └───Extended_1
    └───Extended_2
    └───Extended_3
    └───Extended_5
    └───Extended_7
    └───Extended_8
    └───Figure_1
    └───Figure_2
    └───Figure_4
    └──qc_annotation

5. download necessary data from OSF into a directory named /data within this git repo. Files marked with * must be downloaded to recapitulate analysis steps only. Some files can also be generated, as described in the Description/Origin column. This directory includes the following files:

Data File	Description / Origin
pathway_databases/GO_Biological_Process_2018.txt	from mayaan lab here
pathway_databases/HumanCyc_2016.txt	from mayaan lab here
pathway_databases/KEGG_2019_Human.txt	from mayaan lab here
pathway_databases/Reactome_2016.txt	from mayaan lab here
iPSC_data/FPKM_table_AST.txt	FPKM normalized counts from GEO accession number: GSE102956
iPSC_data/FPKM_table_MIC.txt	FPKM normalized counts from GEO accession number: GSE102956
iPSC_data/FPKM_table_NEU.txt	FPKM normalized counts from GEO accession number: GSE102956
iPSC_data/FPKM_table_OPC.txt	see methods section "Bulk RNA-sequencing from isogenic iPSC-derived oligodendroglia" in our paper; run ../scripts/get_opc_ipsc_counts_table.r
iPSC_data/opc_ipsc_bulk_rnaseq_count_files/	see methods section "Bulk RNA-sequencing from isogenic iPSC-derived oligodendroglia" in our paper
single_cell_data/Cell_group_colors.rds *	colors assigned manually
single_cell_data/expressed_genes_per_celltype.rds	run ./scripts/get_expressed_genes_per_celltype.r
single_cell_data/Mapping.rds *	colors assigned manually
single_cell_data/RefCellTypeMarkers.adultBrain.rds *	Reference cell type marker genes were obtained from PsychENCODE, reported in Wang, D. et al. Comprehensive functional genomic resource and integrative model for the human brain. Science 362, (2018)
single_cell_data/PanglaoDB.by.organ.by.celltype.rds *	downloaded from Panglao database (https://panglaodb.se/); Franzén, O., Gan, L.-M. & Björkegren, J. L. M. PanglaoDB: a web server for exploration of mouse and human single-cell RNA sequencing data. Database 2019, (2019).
differentially_expressed_genes/E4_nebula_associations_Oli.rds	run ./scripts/nebula_degs.r
differentially_expressed_genes/oli_wilcox_results.rds	run ./scripts/get_wilcox_degs.r
differentially_expressed_genes/oli_wilcox_results_AD.rds	run ./scripts/get_wilcox_degs.r
differentially_expressed_genes/oli_wilcox_results_noAD.rds	run ./scripts/get_wilcox_degs.r
differentially_expressed_genes/OPC_deg_statistics.txt	see methods section "Bulk RNA-sequencing from isogenic iPSC-derived oligodendroglia" in our paper
single_cell_data/ensembl.GRCh38p12.genes.complete.annot.rds
supplementary_tables/ *	pathway renaming and selection for figure 1 (manual) & other supplementary tables from the paper

6. Download the single-cell- and lipidomic-related data from Synapse and add these data to the ./data directory according to the directories given in the table below. Files marked with * must be downloaded to recapitulate analysis steps only. Some files can also be generated, as described in the Description/Origin column.

Data File	Description / Origin
single_cell_data/individual_level_averages_per_celltype/Ast.csv	run ./scripts/get_individual_level_averages_object.r
single_cell_data/individual_level_averages_per_celltype/Ex.csv	run ./scripts/get_individual_level_averages_object.r
single_cell_data/individual_level_averages_per_celltype/In.csv	run ./scripts/get_individual_level_averages_object.r
single_cell_data/individual_level_averages_per_celltype/Mic.csv	run ./scripts/get_individual_level_averages_object.r
single_cell_data/individual_level_averages_per_celltype/Oli.csv	run ./scripts/get_individual_level_averages_object.r
single_cell_data/individual_level_averages_per_celltype/Opc.csv	run ./scripts/get_individual_level_averages_object.r
single_cell_data/qc_counts_data/qc_column_metadata.csv *	collected and shared by ROSMAP (Dr David Bennett and colleagues)
single_cell_data/qc_counts_data/qc_counts.mtx *	see methods sections "Quality control for cell inclusion" and "Clustering analysis and QC filtering" in our paper, and ../scripts/qc_and_annotation.r
single_cell_data/qc_counts_data/qc_gene_names.txt *	see methods section "snRNA-seq data preprocessing" in our paper
single_cell_data/raw_counts_data/column_metadata.csv	collected and shared by ROSMAP (Dr David Bennett and colleagues)
single_cell_data/raw_counts_data/gene_names.csv	see methods section "snRNA-seq data preprocessing" in our paper
single_cell_data/raw_counts_data/raw_counts.mtx	see methods section "snRNA-seq data preprocessing" in our paper
cc_lipidomics/Lipidomics_RawData.csv *	see methods section "Untargeted lipidomics of post-mortem corpus callosum" in our paper
cc_lipidomics/Lipidomics_RawData_2.csv *	see methods section "Untargeted lipidomics of post-mortem corpus callosum" in our paper
pfc_lipidomics/ChE_summary_cyc_05312022_all_samples.csv *	see methods section "Untargeted Lipidomics on post-mortem prefrontal cortex" in our paper
pfc_lipidomics/ROSMAP_Lipidomics_Emory_biospecimen_metadata.csv *	accession through Synapse here
pfc_lipidomics/metadata_PFC_all_individuals_092520.tsv	Contact us for this file or Use this file, or use Supplementary Table S8, Tab 2 as input

7. create an empty directory in ./data titled other_analyses_outputs and an empty directory in .data titled supplementary_tables

8. Now run the following code snippets to recapitulate the analysis:

Navigate to the scripts folder, run this first:

conda activate apoe4myelin_env
Rscript ../scripts/generate_qc_sce_object.r 
Rscript ../scripts/get_individual_level_averages_object.r 
Rscript ../scripts/make_metadata_file.r 
Rscript ../scripts/get_pathways.r
Rscript ../scripts/nebula_degs.r 
Rscript ../scripts/get_expressed_genes_per_celltype.r 
Rscript ../scripts/get_opc_ipsc_counts_table.r 

Figure 1

conda activate apoe4myelin_env
Rscript ../scripts/pathway_analyses.r 
Rscript ../scripts/get_figure_1_plots.r

Figure 2

conda activate apoe4myelin_env
Rscript ../scripts/dissecting_cholesterol_dysregulation.r
Rscript ../scripts/lipidomic_analysis_cc.r 
Rscript ../scripts/get_figure_2_plots.r  

Figure 4

conda activate apoe4myelin_env
Rscript ../scripts/get_wilcox_degs.r
Rscript ../scripts/get_figure_4_plots.r

Extended Data Figure 1

conda activate apoe4myelin_env
Rscript ../scripts/plots_for_extended_data_figure_1.r 

Extended Data Figure 2

conda activate apoe4myelin_env
Rscript ../scripts/fgsea_analysis.r
Rscript ../scripts/pseudo_bulk.r
Rscript ../scripts/plots_for_extended_data_figure_2.r 
Rscript ../scripts/e4_effects_stratification_by_AD.r
Rscript ../scripts/e4_stratification_plots.r

Extended Data Figure 3

conda activate apoe4myelin_env
Rscript ../scripts/lipidomic_analysis_pfc_get_data.r
Rscript ../scripts/lipidomic_analysis_pfc_make_plots.r

Extended Data Figure 5

conda activate apoe4myelin_env
Rscript ../scripts/comparison_of_ipsc_and_brain_get_scaled_matrices.r
Rscript ../scripts/comparison_of_ipsc_and_brain_plots.r
Rscript ../scripts/APOE_expression_oligodendrocytes.r
Rscript ../scripts/apoe_expression_ipsc.r  

Extended Data Figure 7

conda activate apoe4myelin_env
Rscript ../scripts/get_postmortem_er_stress_pathways.r
Rscript ../scripts/er_postmortem_plots.r  
Rscript ../scripts/ipsc_gene_perturbations.r  

Extended Data Figure 8

conda activate apoe4myelin_env
Rscript ../scripts/get_wilcox_myelin_plots.r

Or, run the full analysis pipeline:

conda activate apoe4myelin_env
chmod +x run_all_analyses.sh
../scripts/run_all_analyses.sh

If you use these data, please cite

Blanchard, J.W., Akay, L.A., Davila-Velderrain, J., von Maydell, D., et al. APOE4 impairs myelination via cholesterol dysregulation in oligodendrocytes. Nature (2022). https://doi.org/10.1038/s41586-022-05439-w

If you have any questions, notice any inconsistencies, need help, or would like to brainstorm future collaborations and ideas, please don't hesitate to reach out: djuna@mit.edu, jose.davila@fht.org