Performance updates & prevent stats crash (7b3969bb) · Commits · github_fork / ZUMIs

UMIstuffFUN.R

+5 −10

Original line number	Diff line number	Diff line
		@@ -53,17 +53,12 @@ reads2genes <- function(featfiles,chunks,rgfile,cores){
		samcommand<-paste("cat freadHeader; samtools view -x NH -x AS -x nM -x HI -x IH -x NM -x uT -x MD -x jM -x jI -x XN -x XS -R",rgfile,"-@",cores)

		if(length(featfiles)==1){
		reads<-data.table::fread(paste(samcommand,featfiles), na.strings=c(""),
		select=c(12,13,14),header=T,fill=T,colClasses = "character")[
		, c("RG","UB","GE"):=list(.rmRG(V12),.rmUB(V13),.rmXT(V14))
		][,c("V12","V13","V14"):=NULL]
		reads<-data.table::fread(paste(samcommand,featfiles[1],"\| cut -f12,13,14 \| sed 's/RG:Z://' \| sed 's/UB:Z://' \| sed 's/XT:Z://' "), na.strings=c(""),
		select=c(1,2,3),header=T,fill=T,colClasses = "character" , col.names = c("RG","UB","GE") )
		}else{
		reads<-data.table::fread(paste(samcommand,featfiles[1]), na.strings=c(""),
		select=c(12,13,14),header=T,fill=T,colClasses = "character" )[
		, c("RG","UB","GE"):=list(.rmRG(V12),.rmUB(V13),.rmXT(V14))
		][,c("V12","V13","V14"):=NULL][
		,"tmp":=fread(paste(samcommand,featfiles[2]),select=14,header=T,fill=T,na.strings=c(""),colClasses = "character")
		][ ,"GEin":=.rmXT(tmp) ][ ,tmp:=NULL
		reads<-data.table::fread(paste(samcommand,featfiles[1],"\| cut -f12,13,14 \| sed 's/RG:Z://' \| sed 's/UB:Z://' \| sed 's/XT:Z://' "), na.strings=c(""),
		select=c(1,2,3),header=T,fill=T,colClasses = "character" , col.names = c("RG","UB","GE") )[
		,"GEin":=fread(paste(samcommand,featfiles[2],"\| cut -f13,14 \| sed 's/XT:Z://'"),select=2,header=T,fill=T,na.strings=c(""),colClasses = "character")
		][ ,"ftype":="NA"
		][is.na(GEin)==F,ftype:="intron"
		][is.na(GE)==F, ftype:="exon"

statsFUN.R

+25 −30

Original line number	Diff line number	Diff line
		@@ -11,13 +11,9 @@ sumstatBAM <- function(featfiles,cores,outdir,user_seq,bc,outfile){
		write(headerXX,paste(outdir,"freadHeader",sep="/"))
		samcommand<-paste("cat freadHeader; samtools view -x NH -x AS -x nM -x HI -x IH -x NM -x uT -x MD -x jM -x jI -x XN -x UB -@",cores)
		#issue with BC matching
		mapCount<-data.table::fread(paste(samcommand,featfiles[1]), na.strings=c(""),
		select=c(12,13,14),header=T,fill=T,colClasses = "character" )[
		, c("RG","XS","GE"):=list(.rmRG(V12),.rmXS(V13),.rmXT(V14))
		][,c("V12","V13","V14"):=NULL
		][,"tmp":=fread(paste(samcommand,featfiles[2]),select=14,header=T,fill=T,na.strings=c(""),colClasses = "character")
		][ ,"GEin":=.rmXT(tmp)
		][ ,tmp:=NULL
		mapCount<-data.table::fread(paste(samcommand,featfiles[1],"\| cut -f12,13,14 \| sed 's/RG:Z://' \| sed 's/UB:Z://' \| sed 's/XT:Z://' "), na.strings=c(""),
		select=c(1,2,3),header=T,fill=T,colClasses = "character" , col.names = c("RG","UB","GE") )[
		,"GEin":=fread(paste(samcommand,featfiles[2],"\| cut -f13,14 \| sed 's/XT:Z://'"),select=2,header=T,fill=T,na.strings=c(""),colClasses = "character")
		][ ,"ftype":="NA"
		][is.na(GEin)==F,ftype:="Intron"
		][is.na(GE)==F ,ftype:="Exon"
		@@ -129,4 +125,3 @@ totReadBoxplot<-function(typeCount,fillcol){
		axis.title.x=element_blank())
		return(box)
		}

		No newline at end of file

zUMIs-stats2.R

+1 −1

Original line number	Diff line number	Diff line
		@@ -14,7 +14,7 @@ library(cowplot)
		##########################
		myYaml<-commandArgs(trailingOnly = T)
		opt <-read_yaml(myYaml)

		setwd(opt$out_dir)
		featColors<-c("#1A5084", "#914614" ,"#118730","grey33","tan1","gold1","grey73","firebrick3")
		names(featColors)<-c("Exon","Intron+Exon","Intron","Unmapped","Ambiguity","Intergenic","Unused BC","User")
		#####################################

zUMIs.yaml

+1 −4

Original line number	Diff line number	Diff line
		@@ -81,7 +81,4 @@ use_SLURM: no
		which_Stage: Filtering

		#below, the fqfilter will add a read_layout flag defining SE or PE
		read_layout:
		read_layout:
		read_layout:
		read_layout:

Admin message