Skip to content

Latest commit

 

History

History
205 lines (147 loc) · 6.68 KB

README.md

File metadata and controls

205 lines (147 loc) · 6.68 KB
Raw

SCHNAPPsContributions

some contributions to the SCHNAPPs project (https://c3bi-pasteur-fr.github.io/UTechSCB-SCHNAPPs/)

Trajectory related contributions

Tempora: cell trajectory inference using time-series single-cell RNA sequencing data

https://github.com/BaderLab/Tempora https://www.baderlab.org/Software/Tempora

Tempora is a novel cell trajectory inference method that orders cells using time information from time-series scRNAseq data. Tempora uses biological pathway information to help identify cell type relationships and can identify important time-dependent pathways to help interpret the inferred trajectory.

scorpius

https://github.com/rcannood/SCORPIUS https://www.biorxiv.org/content/10.1101/079509v2

SCORPIUS an unsupervised approach for inferring linear developmental chronologies from single-cell RNA sequencing data.

ELPIGraph Elastic principal graphs

https://github.com/Albluca/ElPiGraph.R

The mapping of a principal graph into multidimensional data space is regularized by minimizing the stretching of the graph edges and the deviation from harmonicity for the graph stars.

CelliD

https://github.com/RausellLab/CelliD

Gene signature extraction and cell identity recognition at the single-cell level with Cell-ID, Nature Biotechnology 2021

CelliD is a robust statistical method that performs gene signature extraction and functional annotation for each individual cell in a single-cell RNA-seq dataset.

Self-organizing Maps (SOM) based on the Rsomoclu, and kohonen packages.

Unsupervised clustering of genes.

scDEA

https://github.com/Zhangxf-ccnu/scDEA

https://pubmed.ncbi.nlm.nih.gov/34571530/

applies ensemble learning for single-cell differential expression analysis on single-cell RNA-seq dataset.

This is actually implemented in the main up, but only available if scDEA is installed. Since it is very heavy computationally it is documented only here.

Prepare data for SCHNAPPs

read data from 10X experiments

load some utility functions.

source("functions.R")


seuratList = list() # list of count objects coming from the modified Seurat H5 reader

sampleNames <- c("Buffercells", "GFPcells")

# PROJECTFOLDER = 10X project folder that holds the 'counts' directory
# we are reading filtered H5 files.

for (sIdx in 1:2){
  h5file <- paste0("PROJECTFOLDER/counts/",sampleNames[sIdx],"/outs/filtered_feature_bc_matrix.h5")
  seuratList[[sIdx]] <- myRead10X_h5(h5file, use.names = TRUE, unique.features = TRUE, sampleName = sampleNames[sIdx])
}
scexSeurat = seuratList[[1]]
for (sIdx in 2){
  scexSeurat <- cbind(scexSeurat, seuratList[[sIdx]])
}
# print the result
scexSeurat

# filter genes with no expression
scexSeurat = scexSeurat[rowSums(as.matrix(scexSeurat))>0,]
# make sure that no genes exist with the same overall expression. This would cause problems for some calculations.
scexSeurat = unique(as.matrix(scexSeurat))



scEx <- SingleCellExperiment(
  assay = list(counts = scexSeurat),
  colData = pd,
  rowData = featuredata
)





featureData_summary <- data.frame(
  "Description" = NA,
  "gene_id" = my_gene$gene_id,
  "Chromosome.Name" = my_gene$seqid,
  "Associated.Gene.Name" = my_gene$gene_name,
  stringsAsFactors = F
)
# featureData_summary <- featureData_summary[, -2]
featuredata <- data.frame(featureData_summary, stringsAsFactors = F)
featuredata$id <- featuredata$Associated.Gene.Name
featuredata$symbol <- make.unique(featuredata$Associated.Gene.Name)
featuredata <- featuredata[which(featuredata$symbol %in% rownames(scexSeurat)), ]
featuredata <- unique(featuredata)
featuredata[which(duplicated(featuredata$symbol)), ]

# rownames(featuredata) = 1:nrow(featuredata)
# featureDatatmp <- featuredata[, -2]
# 
# featureDatatmp <- unique(featureDatatmp)

#same symbol, different ensg numbers:
# featuredata = featuredata[rownames(featureDatatmp), ]
rownames(featuredata) <- featuredata$symbol
nrow(featuredata)
nrow(scexSeurat)
nrFD = nrow(featuredata)
newRowNames = rownames(scexSeurat)[which(!rownames(scexSeurat) %in% featuredata$Associated.Gene.Name)]

if (length(newRowNames) > 0) {
  featuredata[(nrFD + 1):(nrFD + length(newRowNames)),] <- NA
  rownames(featuredata[(nrFD + 1):(nrFD + length(newRowNames)),]) <- newRowNames
  featuredata[(nrFD + 1):(nrFD + length(newRowNames)),"symbol"] <- newRowNames
  featuredata[(nrFD + 1):(nrFD + length(newRowNames)),"Description"] <- newRowNames
  featuredata[(nrFD + 1):(nrFD + length(newRowNames)),"Associated.Gene.Name"] <- newRowNames
  featuredata[(nrFD + 1):(nrFD + length(newRowNames)),"id"] <- newRowNames
}

rownames(scexSeurat)[which(!rownames(scexSeurat) %in% featuredata$Associated.Gene.Name)]
rownames(scexSeurat)[which(!featuredata$Associated.Gene.Name %in% rownames(scexSeurat)) ]
# scexSeurat <- scexSeurat[featuredata$Associated.Gene.Name, ]
featuredata <- featuredata[rownames(scexSeurat), ]
# featuredata$Description = featureData_summary93[featuredata$gene_id,"Description"]
nrow(featuredata)
nrow(scexSeurat)


pd <- data.frame(
  barcode = sub("(.*)-(.*)", "\\1", colnames(scexSeurat)),
  sampleNames = sub(".*-(.*)", "\\1", colnames(scexSeurat))
)
pd$barcode <- as.character(pd$barcode)
rownames(pd) <- colnames(scexSeurat)

hs.pairs <- readRDS("mouse_cycle_markers.rds")
ensembl <- mapIds(org.Mm.eg.db, keys=rownames(featuredata), keytype="SYMBOL", column="ENSEMBL")
assignments <- cyclone(scexSeurat, hs.pairs, gene.names=ensembl, BPPARAM = MulticoreParam())
pd$phases = assignments$phases
pd$G1score = assignments$normalized.scores$G1
pd$Sscore = assignments$normalized.scores$S
pd$G2Mscore = assignments$normalized.scores$G2M

pd$phases[is.na(pd$phases)] = "NA"
pd$G1score[is.na(pd$G1score)] = "NA"
pd$Sscore[is.na(pd$Sscore)] = "NA"
pd$G2Mscore[is.na(pd$G2Mscore)] = "NA"


# scExTemp <- applySingleR(scEx, DatabaseImmuneCellExpressionData(), "cellTypes")

# colData(scEx)$cellTypes = as.factor(colData(scExTemp)$cellTypes)

fname = "Ferdinand-scRNAseq"
outfile <- paste0(fname, ".RData")
save(file = outfile, list = c("scEx"))

# set.seed(1)
# colIdx <- sample(1:ncol(scexSeurat), 2000, replace = FALSE)
# scEx <- SingleCellExperiment(
#   assay = list(counts = scexSeurat[, colIdx]),
#   colData = pd[colIdx, ],
#   rowData = featuredata
# )
# outfile <- paste0(fname, ".sml.RData")
# save(file = outfile, list = c("scEx"))

require(rmarkdown)
knitr::opts_chunk$set(
  message = FALSE,
  warning = FALSE,
  echo = FALSE,
  include = TRUE
)
getwd()
rm("params")
render("gbmReport.Rmd",
       output_file = paste0(fname, ".report.html"),
       output_format = "html_document",
       params = list(
         fileN = paste0(fname, ".RData"),
         min.genes = 2,
         min.cells = 3,
         low.thres1 = 2,
         low.thres2 = -Inf,
         high.thres1 = 2500,
         high.thres2 = 2000
       )
)