Skip to content

Latest commit

 

History

History
101 lines (75 loc) · 2.64 KB

README.md

File metadata and controls

101 lines (75 loc) · 2.64 KB
Raw

scExtras

Provides additional functions for Seurat v3

  • Pipeline tools
  • Diffusionmap
  • Slingshot
  • Ligand-Receptor Analysis

Requirements

devtools::install_github('chris-mcginnis-ucsf/DoubletFinder')
devtools::install_github("jokergoo/ComplexHeatmap")
devtools::install_github("Morriseylab/ligrec")

install.packages(c("Seurat","NMF","data.table","broom","quantreg","gam","parallelDist"))

if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("kstreet13/slingshot")
BiocManager::install(c("destiny","scds","Biobase"))

Install

devtools::install_github('Morriseylab/scExtras')

Run Pipeline

Set all input parameters

outdir<-'Seurat' 
projectname<-'HumanLung' 
input10x <- c('STARSolo/HumanLungSolo.out/Gene/filtered/') 

org<-'mouse'
npcs<-40 
k=30

dir.create(outdir,recursive = T,showWarnings = F)
plotdir <- paste0(outdir,'/plots')
dir.create(plotdir,showWarnings = F)
qcdir <-paste0(outdir,'/qc')
dir.create(qcdir,showWarnings = F)

Read in 10x files, this can be H5 files or directories containg the mtx, barcode and gene files and create Seurat Object.

scrna = CreateSeuratObj(files=input10x,name=projectname)

Filter the data and run doublet detector.Doublet detection is performed using scds and DoubletFinder

scrna = RunQC(scrna,org=org,filter = T,LowerFeatureCutoff=200,UpperFeatureCutoff="MAD",UpperMitoCutoff=5,doubletdetection = T,dir=qcdir)

Normalize data and scale data, ccscale=T will regress out cell cycle effects. To run scTransform, use the vignette

scrna = processExper(scrna,ccscale=F,return_var_genes = F,org=org)

Perform PCA and save plots in QC dir.

scrna= PCATools(scrna, npcs=npcs, jackstraw=T, plotdir = qcdir)

Perform Louvain Clustering and UMAP reduction

npcs= 15 #set number of PC's to use based on elbow and jackstraw plots
scrna <- ClusterDR(scrna,dims=1:npcs,n.neighbors =20,min.dist=0.2, findallmarkers =F,res=0.4 )

Diffusion Map

Seruat v3 removed the Diffusionmap dimension reduction routine.

scrna <- RunDiffusion(scrna, dims=1:20)

Trajectory Analysis using slingshot

scrna <- runSlingshot(mes,reduction='umap',approx_points = 200,extend= "n",stretch=0)
lineageDimPlot(scrna,reduction = "umap",group.by = "var_cluster",lineage = "all")

Basic Ligand-Receptor Analysis

scrna <- RunLigRec(scrna,org=org)

##Save seurat object as RDS file

saveRDS(scrna,paste0(outdir,'/',projectname,'.RDS'))