scTPA is a web tool for single-cell transcriptome analysis and annotation based on biological pathway activation in human and mice. We collected a large number of biological pathways with different functional and taxonomic classifications, which facilitates the identification of key pathway signatures for cell type annotation and interpretation.

• Calculating pathway activity score of single cell
• Dimension reduction
• Clustering of cell population by different methods
• Identifying significantly activated pathways of cell clusterings
• Comparison analysis of the associated gene expression profiles of pathways

• step1 Download scTPA scTPA can be download directly from Download ZIP button. Alternatively, scTPA can be installed through github: enter the directory where you would like to install scTPA and run

git clone https://github.com/yupenghe/methylpy.git
cd scTPA/

• step2 Install dependent R packages For using scTPA, user must install following packages,:

Seurat foreach bigstatsr data.table dplyr scales ggplot2 cowplot pheatmap

To install this packages, start "R" and enter:

if (!requireNamespace(c("Seurat","bigstatsr","data.table","foreach","dplyr","scales","ggplot2","cowplot","pheatmap"), quietly = TRUE))
    install.packages(c("Seurat","bigstatsr","data.table","foreach","dplyr","scales","ggplot2","cowplot","pheatmap"))

• step3 Install optional R packages If user want to use some specialized method in scTPA, the following R packages are required.

1. scran for "scran" normalization method
2. scImpute for "scImpute" imputation method
3. SIMLR for "simlr" clustering method
4. dbscan for "dbscan" clustering method

scran

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
if (!requireNamespace("scran", quietly = TRUE))
    BiocManager::install("scran")

scImpute

if (!requireNamespace("devtools", quietly = TRUE))
    install.packages("devtools")
if (!requireNamespace("scImpute", quietly = TRUE))
    devtools::install_github("Vivianstats/scImpute")

SIMLR

if (!requireNamespace("devtools", quietly = TRUE))
    install.packages("devtools")
if (!requireNamespace("scImpute", quietly = TRUE))
    devtools::install_github("Vivianstats/scImpute")

dbscan

if (!requireNamespace("dbscan", quietly = TRUE))
    BiocManager::install("dbscan")

Rscript /path/to/you/scTPA-master/R/scTPA.R -f /path/to/you/scTPA-master/test/expression.csv --cellType /path/to/you/scTPA-master/test/cell_type.csv --work_dir /path/to/you/scTPA-master/ --species homo --pathway_database kegg --para_size 1 -o /path/to/you/scTPA-master/results/

Once the program has run successfully, a series of results files and folders will appear in the results folder.

Rscript /path/to/you/scTPA/scTPA.R -h
Options:
    -f FILE, --file=FILE
        gene expression profile, genes X cells
    --cellType=CELLTYPE
        cell type file. First column is cell name (same as the colnames of gene expression profile), second column is cell type. No header names.[default= NULL]
    --work_dir=WORK_DIR
        Workshop direction. [default= ./]
    --normalize=NORMALIZE_METHOD
        methods used for normalization. "log", "CLR", "RC" or "scran"[default= none]
    --min_cells=MIN_CELLS
        genes must be in a minimum number of cells. Used for filtering genes[default= 3]
    --min_features=MIN_FEATURES
        cells must have at least the minimum number of genes. Used for filtering cells[default= 200]
    --species=SPECIES
        species. "homo" or "mus"[default= homo]
    --imputation=IMPUTATION
        Imputation method. "scImpute" or "none"[default= none]
    --data_type=FILE
        data type of gene expression profile，"TPM" or "count"[default= TPM]
    --pathway_database=PATHWAY_DATABASE
        pathway database, detials see https://github.com/sulab-wmu/scTPA[default= kegg]
    --user_pathway=USER_PATHWAY
        user defined pathway file，only for gmt format[default = NULL]
    --pas_method=PAS_METHOD
        method for calculating PAS. "gsva", "ssgsea", "zscore" or "plage"[default= ssgsea]
    --para_size=PARA_SIZE
        number of kernels used for parallel[default= 4]
    --cluster_method=CLUSTER_METHOD
        clustering method. "seurat", "hclust", "simlr", "kmedoids", "kmeans" or "dbscan"[default= seurat]
    --seurat_dims=SEURAT_DIMS
        dimensions used in Seurat clustering[default= 8]
    --seurat_resolution=SEURAT_RESOLUTION
        resolution used for Seurat clustering[default= 0.5]
    --k_cluster=K_CLUSTER
        number of clusters, useless if clustering method is Seurat or dbscan[default= 5]
    --min_pts=MIN_PTS
        parameter in DBSCAN[default= 3]
    --dims=DIMS
        number of PCA dimensions used for TSNE or UMAP[default= 20]
    --marker_method=FIND_MAKER_METHOD
        method of finding siginificant markers[default= wilcox]
    --logFC_thre=THRESHOLD_LOGFC
        threshold of logFC (Detail see Seurat)[default= 0.25]
    --min_pct=MIN_PCT
        only test genes that are detected in a minimum fraction of min.pct cells in either of the two populations.[default= 0.1]
    --pic_type=PIC_TYPE
        type of picture, png or pdf [default= png]
    -o OUT_DIR, --out_dir=OUT_DIR
        output folder[default= NULL]
    -h, --help
        Show this help message and exit

--normalize: log: Log transform. Feature counts output for each cell is divided by the total counts for that cell and multiplied by 1e4. This is then natural-log transformed. CLR: Centered log ratio. A commonly used Compositional Data Analysis (CoDA) transformation method. RC: Relative counts. Feature counts output for each cell are is divided by the total counts for that cell and multiplied by 1e4 (for TPM/CPM/FPKM/RPKM this value is 1e6). scran: The normalization strategy for scRNA-seq is implemented based on the deconvolutional size factor using the scran R package. Detials see scran none: Do not implement normalization

--imputation: scImpute: Imputing missing value of data matrix following filtering and normalization steps and this function is performed using scImpute R package none: Do not implement imputation.

--data_type: count: Discrete Data. TPM: Continuous data.

--pathway_database:

when "--species" is "homo", "--pathway_database" can be select as follow: kegg: An encyclopaedia for genes reaction and regulation. KEGG. reactome: A curated database for biomolecular pathways. Reactome. biocarta: A pathway database for gene regulation. BioCarta. smpdb: A small molecules pathway database. SMPDB. humancyc: A curated pathway database of human metabonomics. HumanCyc. panther: A curated pathway database for protein annotation through evolutionary relationship. PANTHER. pharmgkb: A curated pathway database for pharmacogenomics. pharmGKB. acsn2: A web-based resource depicting signalling and regulatory molecular processes in cancer cell and tumor microenvironment. ACSN v2.0. rb: A curated map of molecular interactions about retinoblastoma protein (RB/RB1). RB-Pathways. h.all: Hallmark gene sets. MSigDB. c2.cgp: Chemical and genetic perturbations. MSigDB. c2.cp: c4.cgn: Cancer gene neighborhoods. MSigDB. c4.cm: Cancer modules. MSigDB. c5.bp: GO biological process. MSigDB. c5.mf: GO cellular component. MSigDB. c5.cc: GO molecular function. MSigDB. c6.all: Oncogenic signatures. MSigDB. c7.all: Immunologic signatures. MSigDB.

when "--species" is mus, "--pathway_database" can be select as follow: kegg: An encyclopaedia for genes reaction and regulation. KEGG. reactome: A curated database for biomolecular pathways. Reactome. smpdb: A small molecules pathway database. SMPDB. c5.bp: GO biological process. GSKB. c5.mf: GO cellular component. GSKB. c5.cc: GO molecular function. GSKB. other: Including "Location", "HPO", "STITCH", "MPO", "T3DB", "PID", "MethyCancer" and "MethCancerDB*, details see table. GSKB.