sciNOME is an R package for jointly analyzing transcriptomic, DNA methylation, and chromatin accessibility data. The package is designed to process sequencing data from scNOMe-seq and scRNA-seq experiments. Additionally, sciNOME converts sequencing data into region-based formats for efficient storage and subsequent analysis. The R package incorporates features such as differential analysis, dimensionality reduction analysis, and differential analysis of horizontal sites.
# Development version from GitHub
BiocManager::install("Medinfo-Lab/sciNOME")
# From Bioconductor (once accepted)
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("sciNOME")
If you want to install offline, please follow the steps below:
Download the latest sciNOME release from our GitHub Releases page to your local machine.
# Install prerequisite packages
if (!requireNamespace("BiocManager", quietly = TRUE)) install.packages("BiocManager")
deps <- c("data.table", "dplyr", "ggplot2", "ggpubr", "ggridges", "patchwork",
"Matrix", "irlba", "Rtsne", "uwot", "parallel", "pbapply", "igraph",
"FNN", "NMF", "princurve", "tidyr", "ggrepel", "GenomicRanges",
"IRanges", "S4Vectors", "impute")
install.packages("/path/to/your/download/sciNOME_latest.tar.gz", repos = NULL, type = "source")
The Workflow:
library(sciNOME)
library(dplyr)
Here, we will use the built-in data as an example.
# 1.1 Dynamically obtain the absolute path of test data included in the package
rna_path <- system.file("extdata", "RNA_counts.txt", package = "sciNOME")
meta_path <- system.file("extdata", "metadata.csv", package = "sciNOME")
bed_path <- system.file("extdata", "target_promoters.bed", package = "sciNOME")
cpg_dir <- system.file("extdata/cpg_cov", package = "sciNOME")
gpc_dir <- system.file("extdata/gpc_cov", package = "sciNOME")
# 1.2 Read basic expression profiles, metadata, and target interval files
rna_counts <- read.table(rna_path, sep="\t", header=TRUE)
metadata <- read.csv(meta_path)
bed_data <- read.table(bed_path,header = T)
# 2.1 Build lightweight RNA analysis objects
RNA_obj <- Build_RNAObject(
expr_mat = rna_counts, # Expression Matrix
meta_data = metadata, # Grouped Data Frames
meta_id_col = "RNA_ID", # The columns in the grouped data are used to match the column names in the matrix
min_cells = 0, # (Optional) Remove genes that are completely non-expressed
min_features = 0, # (Optional) Remove completely empty samples
project_name = "RNA_object"
)
PlotQC_RNA(RNA_obj,"CellType")
# 2.2 Data Filtering and Standardization
RNA_obj <- ProcessQC_RNA(
obj = RNA_obj,
mt_pattern = "^MT-", # Human mitochondrial prefix
min_nFeature = 0, # Express at least 0 genes
max_nFeature = 6000, # Up to 6000 genes
min_nCount = 0, # At least 0 UMI
max_nCount = 100000, # Up to 100000 UMI
max_mt = 10, # The proportion of mitochondria must not exceed 10%
norm_method = "LogNormalize", # Standardized Methods
do_scale = TRUE # Whether to scale
)
# 2.3 Dimensionality Reduction Analysis (PCA)
RNA_obj <- RunDimReduction_RNA(
obj = RNA_obj,
method = "PCA",
n_hvg = 2000,
layer_name = "data",
pca_rank = 10,
umap_neighbors = 100,
umap_mindist = 2
)
PlotDimRed_RNA(RNA_obj,"PCA","CellType",show_cluster = F)
# 2.4 Differential Expression Analysis (DEA)
RNA_diff <- RunDEA_RNA(
obj = RNA_obj,
group_col = "CellType",
ident_1 = "CellType.A"
# ident_2 = "CellType.B"
)
PlotVolcano_RNA(RNA_diff,
fc_cut = 0.5,
p_cut = 0.05)
# 3.1 Aggregate apparent site data based on the given interval
result_matrixCpG <- Aggregate_epiRegions(
cov_dir = gpc_dir,
bed_file = bed_path,
n_cores = 8
)
result_matrixGpC <- Aggregate_epiRegions(
cov_dir = gpc_dir,
bed_file = bed_path,
n_cores = 8
)
# 3.2 Missing Value Quality Control and Filtering
qc_result_matrixCpG <- QC_epiData(
data = result_matrixCpG,
top_n_rows = 2000,
max_col_na_ratio = 0.8
)
qc_result_matrixGpC <- QC_epiData(
data = result_matrixGpC,
top_n_rows = 2000,
max_col_na_ratio = 0.8
)
# 3.3 Extract methylation/open chromatin level matrix for downstream dimensionality reduction
qc_result_matrixCpG_level <- Extract_epiData(qc_result_matrixCpG,".level")
qc_result_matrixGpC_level <- Extract_epiData(qc_result_matrixGpC,".level")
# 3.4 Multidimensional Scaling (MDS) for Phenotypic Data
MDS_df_CpG <- Reduce_epiDims(
qc_result_matrixCpG_level, # choose GpC data (option)
metadata,
"CpG_level","CellType",
dr_method = "MDS",
impute_method = "knn"
)
PlotDimRed_Epi(MDS_df_CpG,"CellType","MDS")
# 3.5 Identifying Differentially Methylated Regions (DMRs)
CpG_diff <- Run_Diffanalysis(
raw_mat = qc_result_matrixCpG,
meta = metadata,
group_col = "CellType",
target_group = "CellType.A",
# control_group = "Unt",
col_level = "CpG_level",
col_meth = "CpG_meth",
col_nonmeth = "CpG_nonmeth")
# DMRs plot (option)
PlotVolcano_Epi(CpG_diff,th_effect = 0.05,title = "CellType.A vs other CpG Volcano") # Differential region analysis (DMRs)
# 3.6 Identifying Differentially Accessible Chromatin Regions (DARs)
GpC_diff <- Run_Diffanalysis(
raw_mat = qc_result_matrixGpC,
meta = metadata,
group_col = "CellType",
target_group = "CellType.A",
# control_group = "Unt",
col_level = "CpG_level",
col_meth = "CpG_meth",
col_nonmeth = "CpG_nonmeth")
# DARs plot (option)
PlotVolcano_Epi(GpC_diff,th_effect = 0.05,title = "CellType.A vs other GpC Volcano") # Differential region analysis (DARs)
# 4.1 Multi-omics data integration
integrated_obj <- Integrate_MultiOmics(
mode = "tri", # Integration mode
target_group = "ALL", # Target group
meta_df = metadata, # Grouped Data Frames
group_col = "CellType", # Columns that define groups
region_df = bed_data, # Regional Information Table
rna_obj = RNA_obj, # Transcriptomic objects
rna_id_col = "RNA_ID", # The column name containing the RNA groups
# rna_diff_df = RNA_diff, # RNA differential data (option)
cpg_mat = qc_result_matrixCpG_level, # The level data for each CpG sample
cpg_id_col = "CpG_level", # The column name containing the CpG groups CpG
# cpg_diff_df = CpG_diff, # DNA methylation differential data (option)
gpc_mat = qc_result_matrixGpC_level, # The level data for each GpC sample
gpc_id_col = "GpC_level" # The column name containing the CpG groups GpC
# gpc_diff_df = GpC_diff, # Chromatin accessibility differential data (option)
)
# 4.2 Draw multi-omics level heatmaps / matrix plots
PlotOmicsMatrix(integrated_obj)
# 4.3 Draw multi-omics scatter correlation plots (e.g., Methylation vs Expression vs Accessibility)
PlotOmicsScatter(integrated_obj)
MIT License — see LICENSE for details.