Brain Single-Cell Omics for PsychEncode (brainSCOPE) Output Files

Here are all output files associated with the brainSCOPE Resource, incorporating single cell-resolution data for 388 adult DLPFC samples. Additional details for each data file are located in the Supplementary document for the associated manuscript. Files are organized into eight sections according to the manuscript section headings, and are fully open source. Citations are listed for files included as Supplementary data in the manuscript (i.e. Data S1).

A select list of Key Resources suitable for most downstream analyses is also available here.

Constructing a single-cell genomic resource for 388 individuals

26K	.txt	PEC2_sample_metadata.txt	Clinical and demographic meta-data for each sample in the brainSCOPE resource. [Data S1]
19K	.xlsx	PEC2_sample_mapping.xlsx	Mapping of uniform IDs for each sample acorss sub-cohorts and data modalities (snRNA-Seq, snATAC-Seq, and genotype data). [Data S2]
8.1G	.txt	[sample]-annotated_matrix.txt.gz	Expression matrices for individual samples in our cohort, arranged into sub-directories by cohort. (Note that expression matrices for ROSMAP cohort samples are available only through the AMP-AD Knowledge portal.)
710M	.BED	[celltype].expr.bed.gz	Pseudo-bulk snRNA-Seq expression matrices for 24 cell types, listing logCPM normalized expression values for subsets of 388 individuals who pass quality control for each cell type. Data used as input for the scQTL analysis.
46K	.tar	PsychENCODE_scRNA_pipeline-main.tar.gz	GZIP file containing code used for processing snRNA-Seq datasets. More details are available on the associated GitHub repository.
1.7G	.RDS	BICCN_mat.RDS.gz	R Data Object contains the DLFPC cell-type annotation scheme from the BICCN Consortium.
2.1G	.RDS	Ma_Sestan_mat.rds.gz	R Data Object contains the DLFPC cell-type annotation scheme from the Ma-Sestan study.
762K	.RDS	BICCN_meta_share.RDS	R object containing the harmonized DLPFC cell-type annotation scheme used for analysis in the study.
3.9K	.R	Azimuth_mapping.R	R script used to annotate cell types in an Azimuth object.
4.0K	.py	reconcile_annotations.py	Python script used to generate the harmonized cell typing scheme from the input matrices.
817K	.CSV	cell_type_fraction_count_with_meta.csv	Normalized cell fractions calculated from snRNA-Seq data. [Data S4]
120K	.txt	bisque.PEC_CMC.rev.txt	Normalized cell fractions calculated from deconvolved bulk RNA-Seq data. [Data S5]
1.1K	.txt	cell_fraction_corr.txt	Correlations between deconvolved and single-cell derived fractions per cell type. [Data S6]
33M	.csv	ASD_DEGcombined.csv	Sets of differentially-expressed genes between control individuals and individuals with ASD for 20 cell types. [Data S7]
39M	.csv	Bipolar_DEGcombined.csv	Sets of differentially-expressed genes between control individuals and individuals with Bipolar disorder for 19 cell types. [Data S7]
21M	.csv	Schizophrenia_DEGcombined.csv	Sets of differentially-expressed genes between control individuals and individuals with Schizophrenia for 21 cell types. [Data S7]
17M	.csv	*_ASD_table.csv	Different version of differentially-expressed genes between control individuals and individuals with ASD for 20 cell types, used for LNCTP model comparisons.
19M	.csv	*_Bipolar_disorder_table.csv	Different version of differentially-expressed genes between control individuals and individuals with BPD for 19 cell types, used for LNCTP model comparisons.
40M	.csv	*_Schizophrenia_table.csv	Different version of differentially-expressed genes between control individuals and individuals with SCZ for 21 cell types, used for LNCTP model comparisons.
1.0K	.tsv	IT_neuron_trajectory_sig_gene_set.tsv	Overlapping set of genes across cohorts identified as significantly varying (FDR<0.05) across IT neuron trajectories in each cohort. [Table S5]

Determining regulatory elements for cell types from snATAC-seq

43M	.BED	[celltype].PeakCalls.bed	Files contain scCREs (i.e. single-cell candidate cis-regulatory regions) identified from snATAC-Seq and snMultiome data (7 files for individual cell types & 1 file for peaks across all cell types).
67M	.BIGWIG	[celltype].bigwig	Files contain signal tracks for chromatin accessibility derived from samples in each of three snATAC-Seq and snMultiome cohorts assessed (7 files for individual cell types & 1 file for peaks across all cell types).
14M	.bed	adult_bcCREs.bed	List of brain candidate cis-regulatory elements (b-cCREs). These elements can also be visualized on PsychScreen.
347K	.csv	ukbb-all-traits-pval.csv	File containing the -log(p-value) of UKBiobank GWAS summary statistic LDSC enrichment for scCREs in seven cell types and adult b-cCREs regions, indexed by trait ID. [Data S9]
625K	.txt	ccre.pval.table.txt	File containing the -log(p-value) of UKBiobank GWAS summary statistic LDSC enrichment on adult cCREs regions, indexed by trait ID.
6.5K	.csv	PGC_PASS-all-traits-pval.csv	File containing the -log(p-value) of PGC and PASS-derived GWAS summary statistic LDSC enrichment for scCREs in seven cell types and adult b-cCREs regions, indexed by trait ID. [Data S10]
686K	.txt	IDtoTraitName.txt	This file contains a matrix to convert trait ID into the full UKBiobank trait name.
173K	.txt	cluster.trait.txt	File containing a matrix that assigns each UK BioBank trait to a Human Phenotype Ontology category.
156K	.csv	*tf_enrich.csv	Enrichments of TF binding motifs in cell-type-specific scCREs for proximal and distal regions. [Data S11]
190K	.bed	starrseq_enhancers_merged.bed	Lists of enhancer elements validated by STARR-Seq experiments.

Measuring transcriptome and epigenome variation across the cohort at the single-cell level

951K	.csv	variancePartition_output.csv	Summary of VariancePartition results for all genes, including total variation and variation from individual factors (sample and cell type). [Data S12]
1.2M	.csv	variation-partition-092623.csv	Summary of VariancePartition results for all genes, including total variation and variation from individual factors (sample and cell type) plus brain region. [Data S13]
876K	.txt	gene_map_df_02252023.txt	Data frame mapping genes to their gene families (i.e., serotonin). [Data S14]
12M	.bed	gene_coding_conservation_phastcons.bed	Coding gene conservation calculated using phastCons.
9.4M	.bed	bcCRE_conservation_phastcons.bed	Conservation of b-cCREs using phastCons.
34M	.bed	scATAC_conservation_phastcons.bed	Conservation of cell-type-specific snATAC-Seq peaks using phastCons.

Determining cell-type-specific eQTLs from single-cell data

168M	.dat	[celltype]_sig_QTLs.dat	Primary set of cell-type-specific eQTLs identified for 17 cell types, processed with a standardized/conservative pipeline.
45K	.dat	[celltype]_sig_eGenes.dat	Lists of significant eGenes identified for 17 cell types, processed with the standardized pipeline. [Data S15]
201M	.dat	[celltype].SNPs.dat.gz	List of all variant IDs used for scQTL analysis.
8.6M	.bed	[celltype].cov.100_expr_PCs.bed	Sample-level covariates used for QTL calculations in each cell type (i.e., biological sex, age of death, diagnosis, cohort, genotype and expression PCs).
7.7K	.TAR	core_scqtl_processing_code.tar.gz	GZIP file containing code used for calculating scQTLs with the standardized pipeline.
7.3M	.dat	[celltype]_sig_QTLs.dat	Cell-type-specific eQTLS identified for 17 cell types, processed with a standard pipeline utilizing linkage disequilibrium variant selection.
168M	.DAT	[aggregated_celltype]__sig_QTLs.dat	Cell-type-specific eQTLS identified for three PFC cell type groupings (inhibitory, excitatory, and non-neuronal cells), processed with a standard pipeline.
230M	.TXT	conditional.[celltype].txt	eGenes for scQTLS calculated using a conditional QTL identification pipeline.
1.9M	.TXT	conditional_top_variants_per_signal_[celltype].txt	Top variants identified for scQTLs calculated using a conditional QTL identification pipeline.
6.0M	.DAT	bayesian_scQTLs.dat	Set of cell-type-specific QTLs processed with a Bayesian model-based QTL identification pipeline (for recovery of scQTLs in rare celltypes).
48K	.TAR	Bayesian_QTL_code.tar.gz	GZIP file containing code used for calculating scQTLs with a Bayesian model-based pipeline.
1.2M	.TSV	full_dynamic_eqtl.tsv	Full set of eQTLs used as input for the PME model (6225 top eQTLs identified with the pseudo-bulk approach).
797K	.TSV	sig_dynamic_eqtl.tsv	File listing 4186 significant dynamic eQTLs identified with the PME model (SNP term only). [Data S18]
325K	.TSV	sig_ptime_dynamic_eqtl.tsv	File listing the 1,692 significant dynamic eQTLs identified with the PME model (SNP and interaction terms). [Data S19]
21M	.TSV	SZBD.ptime.tsv	This file lists the pseudo-time values for all tested excitatory neuron cells from samples in the SZBDMulti-Seq cohort (generated via SCALEX and Slingshot).
198M	.TXT	metaQTLs.txt	This file lists a combined set of QTLs discovered in all analyses, including analysis type, cell type, gene and SNP identifiers, and p-values and test statistics.
288K	.txt	[celltype]_permuted_fdr_correct.nominal05.txt	List of isoGenes in cell-type-specific isoform usage QTLs for 22 cell types (permuted beta distribution-derived p<0.05, with FDR values for each isoGene).
31M	.txt	[celltype]_significant_sCQTLs_p0.05.txt	isoSNPs for cell-type isoform QTLs, filtered for isoGene-specific nominal p-value thresholds (permuted beta distribution-derived p<0.05 filter) in 22 cell types. [Data S16]
839K	.bed	starrseq_enhancers_merged_sig_qtls_fdr_0.05.bed	Panel design used in mut-STARRseq by intersecting STARRseq validated enhancers with cell-type-specific eQTLs.
41M	.tsv	MultiomeBrain_scASE_genes_anonymized.tsv	List of genes with allele-specific expression in individual cell types, with sample identifiers removed.
38M	.tsv	MultiomeBrain_scASE_hetSNVs_anonymized.tsv	List of heterozygous SNVs conferring allele-specific expression in individual cell types, with sample identifiers removed.
4.6K	.txt	scQTL_disease_overlap.txt	List of eGenes found in any cell type from the primary analysis that are annotated for four brain-related diseases and traits. [Data S17]

Building a gene regulatory network for each cell type

207M	.txt	[celltype]_GRN.txt	Cell-type-specific gene regulatory networks constructed for 24 cell types.
267M	.txt	[celltype].txt	Alternate version of the cell-type-specific gene regulatory networks used as input for the LNCTP model.
143K	.CSV	SCENIC_RSS.csv	SCENIC-derived scores for regulons (TFs with all target genes) in 24 cell types, used as inputs for constructing the final GRNs. [Data S20]
34M	.txt	unified_GRN.txt	Unified gene regulatory network built across all 24 cell types.
1.0M	.txt	[celltype].eqtl_edge.txt	Cell-type-specific lists of scQTLs that overlap with enhancer/promoter elements in the GRNs. [Data S21]
9.5K	.xlsx	crispr_validation_results.xlsx	List of GRN-predicted enhancers validated in targeted CRISPR KO experiments. [Data S22]
629B	.txt	inhubs.txt	Lists of cell-type-specific "in-hub" genes in GRNs. [Data S23]
2.3K	.txt	outhubs.txt	Lists of cell-type-specific "out-hub" genes in GRNs. [Data S23]
945B	.txt	bottlenecks.txt	Lists of cell-type-specific "bottleneck" genes in GRNs. [Data S23]
5.4K	.txt	GO_enrichment_bottlenecks.txt	Gene ontology enrichment among bottleneck genes in cell-type-specific GRNs. [Data S24]
78M	.txt	Unified_GRN_diffusion.txt	List of diffusion scores between TFs and target genes for the unified GRN and each cell-type-specific GRN.
4.4M	.csv	gene_module_mappings.csv	List of module memberships for genes within each cell type-specific GRN.

Constructing a cell-to-cell communication network

1.7M	.txt	cellchat_C2C_network_[disorder].txt	Networks of ligand-receptor signaling patterns across cell types in control, schizophrenia, and bipolar disorder individuals. [Data S25]
320K	.txt	cellchat_C2C_network_netP_[disorder].txt	Networks of ligand-receptor patterns summarized by signaling pathway across cell types in control, schizophrenia, and bipolar disorder individuals. [Data S26]
1.9G	.rds	SZBD-Kellis_annotated-CON_cellchat.rds.gz	CellChat object containing ligand-receptor signaling patterns across cell types in control individuals.
1.6G	.rds	SZBD-Kellis_annotated-SZ_cellchat.rds.gz	CellChat object containing ligand-receptor signaling patterns across cell types in schizophrenia individuals.
1.4G	.rds	SZBD-Kellis_annotated-BD_cellchat.rds.gz	CellChat object containing ligand-receptor signaling patterns across cell types in bipolar disorder individuals.
14G	.h5s	SZBD-Kellis_annotated-BD_CON.h5seurat.gz	NicheNet object containing changes in cell-to-cell networks between individuals with bipolar disorder and control individuals.
15G	.h5s	SZBD-Kellis_annotated-SCZ_CON.h5seurat.gz	NicheNet object containing changes in cell-to-cell networks between individuals with schizophrenia and control individuals.

Assessing cell-type-specific transcriptomic and epigenetic changes in aging

39M	.csv	Aging_DEGcombined.csv	Sets of differentially-expressed genes between older and younger control individuals for 20 cell types.
37M	.csv	Aging-schizophrenia_DEGcombined.csv	Sets of differentially-expressed genes between older and younger individuals with schizophrenia for 20 cell types.
6.0M	.txt	[celltype].txt	Sets of differentially-expressed genes for sample age derived from STEM analysis for 17 cell types.
6.4M	.csv	[celltype]_shap_summary_stratify.csv	Sample age prediction gene priortization results per cell type, derived from the SHAP model.
3.8G	.matrix	bwaob_output_col5_[cell-type].matrix.gz	Deconvolution of bulk ATAC signal to cell-type-specific signals for 7 cell types.
6.8K	.txt	bulk_ATAC_samples_ordered.cleaned.txt	Column (sample) names for the deconvoluted signal matrices.
1.1M	.csv	pec_atacseq_metadata_08142021.csv	Biosample metadata for bulk ATAC-Seq datasets.
1.7K	.txt	hybrid.admodel.auprc.txt	AUPRC values for cell types and data modalities (rf.meth=methylation, rf.expr=expression) from AD model predictions. [Data S29]
9.2G	.tar	shap_xgboost.tar.gz	TAR file containing source code and full results for gene expression prediction model for aging.

Imputing gene expression and prioritizing disease genes across cell types with an integrative model

25K	.XLSX	LNCTP_prioritized_genes_celltypes.xlsx	Tables of genes, cell types, and network elements prioritized by the LNCTP model for each disorder, including salience, coheritability, and p-values for each gene-cell type combination. [Data S30]
17M	.XLS	clue_significant.xls	List of perturbagens identified through CLUE.io that elicit expression profiles opposing perturbations induced by key genes prioritized in LNCTP.
2.1G	.TAR	lnctp.tar.gz	TAR file containing Docker image of source code used to perform the LNCTP model analysis. This is also available to download from the Docker repository.
289M	.TXT	Metacells_Zscores_all.txt.gz	Normalized MetaCell gene expression values for 7 cell types across individuals who pass quality control filters, used as input for GRNs and the LNCTP model.
1.4G	.ZIP	perturbed_expression.zip	Zipped archive of *.csv files containing LNCTP forward perturbation results of key genes identified by LNCTP.
1.4G	.ZIP	perturbed_expression_reversed.zip	Zipped archive of *.csv files containing LNCTP reverse perturbation results of key genes identified by LNCTP.
1.8G	.ZIP	perturbed_expression_background.zip	Zipped archive of *.csv files containing LNCTP forward perturbation results of background genes.
1.8G	.ZIP	perturbed_expression_background_reversed.zip	Zipped archive of *.csv files containing LNCTP reverse perturbation results of background genes.
1.4G	.ZIP	perturbed_expression_drugs.zip	Zipped archive of *.csv files containing LNCTP forward perturbation results of drug target genes.
1.4G	.ZIP	perturbed_expression_drugs_reversed.zip	Zipped archive of *.csv files containing LNCTP reverse perturbation results of drug target genes.
6.6G	.ZIP	perturbation_crispr.zip	Zipped archive of *.csv files containing LNCTP perturbation results of CRISPR targets.
237M	.ZIP	perturbation_crispr_ref.zip	Zipped archive of *.csv files containing reference expression data used in CRISPR perturbation experiments.