Determining cellular heterogeneity in the human prostate with single-cell RNA sequencing

• Contact: Gervaise H. Henry
  • orcid.org/0000-0001-7772-9578
  • Email: gervaise.henry@utsouthwestern.edu
• Institution: UT Southwestern Medical Center
• Department: Urology
• Lab: Strand Lab
  • PI: Douglas W. Strand, PhD
    • orcid.org/0000-0002-0746-927X
    • PI Email: douglas.strand@utsouthwestern.edu
• ANALYZED DATA FOR QUERYING AT: StrandLab.net
• Raw data at: GEO & GenitoUrinary Development Molecular Anatomy Project (GUDMAP)
• Publication at: BioRxiv

Data Analysis

• Requirements:

  • /analysis/DATA/
    • Pd-demultiplex.csv
    • 10x/
      • aggregation_csv.csv
      • GRCh38/
        • "barcodes.tsv"
        • "genes.tsv"
        • "matrix.mtx"
  • 10x CellRanger analyzed data:
    • "filtered_gene_bc_matrice_mex" folder
    • csv file cellranger aggr used for aggregation
    • demultiplex csv file to define subsets of samples
  • R (v3.4.1) packages:
    • methods (v3.4.1)
    • optparse (v1.4.4)
    • Seurat (v2.3.1)
    • readr (v1.1.1)
    • fBasics (v3042.89)
    • pastecs (v1.3.21)
    • qusage (v2.10.0)
    • RColorBrewer (v1.1-2)
    • monocle (v2.6.3)
    • dplyr (v0.7.6)
    • viridis (v0.5.1)
    • and all dependencies

• HOW TO RUN

  • 1 Run on 3 patient aggregate
    • run bash script sc_TissueMapper-Pr.sh
  • 2 Run on 1st patent FACS samples
    • run bash script sc_TissueMapper-D17_FACS.sh
  • 3 Run on 2nd patient FACS samples
    • run bash script sc_TissueMapper-D27_FACS.sh
  • 4 Run on several downsamples from 1 sample from 1st patient
    • run bash script sc_TissueMapper-DS_D17.sh
  • 5 Aggregate and compare several downsamples from #4
    • run bash script sc_TissueMapper_RUN.DS_D17.aggr.R

• Pipeline:

  • Link cellranger count/aggr output to analysis
  • Create demultiplex file to add custom sample groups
  • Load R packages
  • Create analysis folders
  • Load analysis parameters (from default or overwrite from command line)
  • Load cellranger data into R/Seurat
  • Label cells based on their cell cycle stated using Seurat based method
  • QC and filter cells/genes
  • If analyzing samples from multiple patients: Align experiments using canonical correlation analysis (CCA)
  • If analyzing samples from one patient: Perform principle component analysis (PCA) using most highly variable genes (HVG) for downstream clustering etc
  • Perform initial "over" clustering
  • Identify "highly stressed" cells using custom PCA based analysis, remove stressed clusters/cells, and re-cluster
  • Correlate cluster gene expression using Quantitative Set Analysis for Gene Expression (QuSAGE) on lineage genesets for identification (epithelia, and stroma)
  • Subset epithelia from stroma for additional analaysis
    • Re-cluster cell types separately
      • Correlate cluster gene expression using QuSAGE on epithelial subtype genesets for identification (basal, luminal and "other")
      • Correlate cluster gene expression using QuSAGE on stromal subtype genesets for identification (fibroblasts, smooth muscle, endothelia and leukocyte)
    • Optional: Correlate cluster gene expression using QuSAGE on additional genesets for analysis
  • Merge epithelial and stromal cells
  • Identify neuroendocrine cells from epithelial cells using custom PCA based analysis
  • Tabulate population cell numbers
  • Generate differentially expressed genelists (DEGs) of populations

• Genesets:

  • Cell cycle:
    • "regev_lab_cell_cycle_genes.txt" G2M and S phase genes from Genome Res. 2015 Dec; 25(12): 1860–1872
  • Stress:
    • "DEG_C2.CGP.M10970.txt" MSigDB C2 Chemical and Genetic Perturbations M10970 CHUANG_OXIDATIVE_STRESS_RESPONSE_UP
    • "genes.deg.Stress.csv" DWS generated DEGs of stressed cells from scRNA-Seq of 3 patient aggregate
  • Lineage:
    • "DEG_Epi_5FC.txt" DWS generated DEGs of epithelia from FACS population (bulk) RNA-sequencing
    • "DEG_FMSt_5FC.txt" DWS generated DEGs of fibromuscular stroma from FACS population (bulk) RNA-sequencing
    • "genes.deg.Epi.csv" DWS generated DEGs of epithelial cells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.St.csv" DWS generated DEGs of stromal cells from scRNA-Seq of 3 patient aggregate
  • Epithelia:
    • "DEG_BE_5FC.txt" DWS generated DEGs of basal epithelia from FACS population (bulk) RNA-sequencing
    • "DEG_LE_5FC.txt" DWS generated DEGs of luminal epithelia from FACS population (bulk) RNA-sequencing
    • "DEG_OE_5FC.txt" DWS generated DEGs of "other" epithelia from FACS population (bulk) RNA-sequencing
    • "genes.deg.BE.csv" DWS generated DEGs of basal epithelial cells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.LE.csv" DWS generated DEGs of luminal epithelial cells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.OE1.csv" DWS generated DEGs of "other" epithelia cluster 1 cells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.OE2.csv" DWS generated DEGs of "other" epithelia cluster 2 cells from scRNA-Seq of 3 patient aggregate
  • Stroma:
    • "DEG_C5.BP.M11704.txt" MSigDB C5 GO Biological Processes M11704 GO_ENDOTHELIAL_CELL_DIFFERENTIATION
    • "DEG_C5.BP.M10794.txt" MSigDB C5 GO Biological Processes M10794 GO_SMOOTH_MUSCLE_CELL_DIFFERENTIATION
    • "DEG_C5.BP.M13024.txt" MSigDB C5 GO Biological Processes M13024 GO_REGULATION_OF_FIBROBLAST_PROLIFERATION
    • "DEG_C5.BP.M10124.txt" MSigDB C5 GO Biological Processes M10124 GO_LEUKOCYTE_ACTIVATION
    • "genes.deg.Fib.csv" DWS generated DEGs of fibroblast cells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.SM.csv" DWS generated DEGs of smooth muscle cells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.Endo.csv" DWS generated DEGs of endothelial ccells from scRNA-Seq of 3 patient aggregate
    • "genes.deg.Leu.csv" DWS generated DEGs of leukocyte cells from scRNA-Seq of 3 patient aggregate
  • Neuroendocrine:
    • "EurUrol.2005.NE.txt" Neuroendocrine markers from Table 1 of Eur Urol. 2005 Feb;47(2):147-55
    • "genes.deg.NE.csv" DWS generated DEGs of neuroendocrine epithelial cells from scRNA-Seq of 3 patient aggregate
  • Lung epithelia from Lung Gene Expression Analysis (LGEA) Web Portal:
    • "Basal cells-signature-genes.csv" scRNA-Seq LGEA generated top 20 DEGs for [human lung Basal Cells] (https://research.cchmc.org/pbge/lunggens/lungDisease/celltype_IPF.html?cid=3)
    • "Normal AT2 cells-signature-genes.csv" scRNA-Sequecing LGEA generated top 20 DEGs for human lung Alveolar Type 2 Cells
    • "Club_Goblet cells-signature-genes.csv" scRNA-Sequencing LGEA generated top 20 DEGs for human lung Club/Goblet Cells
  • Lung epithelia from Nature 2018 Aug;560(7718):319
    • "SupTab3_Consensus_Sigs.csv" scRNA-Sequencing DEGs of mouse basal, club, ciliated, tuft, neuroendocrine, ionocyte cells (Supplementary Tables, SupTab3_Consensus_Sigs)
    • "SupTab6_Krt13_Hillock.csv" scRNA-Sequencing DEGs of mouse Krt13+ hillock cells (Supplementary Tables, SupTab6_Krt13_Hillock)
    • "Ensemble.mus-hum.txt" Ensemble export of mouse (GRCm38.p6) to human ortholog mapping
  • General MSigDb
    • "c2.all.v6.1.symbols.gmt" MSigDB C2 Curated Gene Sets MSigDB C2
    • "c2.cp.kegg.v6.1.symbols" MSigDB C2 KEGG Gene Subsets KEGG
    • "c5.all.v6.1.symbols.gmt" MSigDB C5 Gene Ontology Gene Sets MSigDB C5
    • "c5.bp.v6.1.symbols.gmt" MSigDB C5 Gene Ontology Biological Processes Gene Subsets GO BP