zUMIs2.7.2 new BC features

zUMIs will try to automatically do this, otherwise convert zUMIs output to loom by simply running `Rscript rds2loom.R myRun.yaml`.

## Changelog

27 Mar 2020: zUMIs2.7.2: New barcode handling functionalities: When using intersection of automatic BC detection and BC whitelist and the full barcode is composed out of several barcode pieces (eg. RT barcode + illumina barcode), the whitelist can now also just be corresponding to just one of the barcode pieces (eg. RT barcode only whitelist). Furthermore, some scRNA-seq protocols may have several cell barcodes that belong to the same cell (eg. SPLiT-seq with oligo-dT/random-hex round 1 barcode; i7 barcode mix in 10x Genomics). zUMIs now supports internally combing the counts via the `barcode_sharing:` option. Please look at the [wiki for further details](https://github.com/sdparekh/zUMIs/wiki/Barcodes#barcode-sharing-feature) and at [examples for some protocols](https://github.com/sdparekh/zUMIs/wiki/Protocol-specific-setup).

16 Mar 2020: zUMIs2.7.1: Smart-seq3 data can be run with the proper consideration of strand information. When setting `strand: 1`, UMI reads will use this strand while non-UMI reads will stay unstranded.

19 Feb 2020: [zUMIs2.7.0 released](https://github.com/sdparekh/zUMIs/releases/tag/2.7.0): Simplify installation greatly by a cond-pack of miniconda with all dependencies. The conda environment is used with `zUMIs-master.sh -c`, with the old behavior staying as the default. 

    .barcode_plot(bccount,outfilename)

  }

  if(length(bccount[,XC] %in% bc_wl)>0){

  if(nchar(bccount[keep==TRUE,XC][1]) != nchar(bc_wl[1])){

    print("length of barcodes not equal to given barcode list, trying to match up...")

    search_vector <-  bccount[keep==TRUE,XC]

    bc_matched <- parallel::mcmapply(function(x) grep(pattern = x, x =search_vector), bc_wl, mc.cores = opt$num_threads, mc.preschedule = TRUE)

    bc_matched <- unlist(bc_matched)

    if(length(bc_matched)>0){

      to_remove <- search_vector[-bc_matched]

      bccount[XC %in% to_remove,keep:=FALSE]

    }

  }else if(sum(bccount[keep==TRUE,XC] %in% bc_wl)>0){

    bccount[ !(XC %in% bc_wl),keep:=FALSE]

  }else{

    warning("None of the frequent barcodes is present in the whitelist. Keep all automatically detected BCs.")

                                                                            !( XC %in% true_BCs )   ]

  nocell_BCs <- nocell_bccount[,XC]

  if(opt$barcodes$BarcodeBinning>0){

    #break up in pieces of 1000 real BCs in case the hamming distance calculation gets too large!

    true_chunks <- split(true_BCs, ceiling(seq_along(true_BCs)/1000))

    for(i in 1:length(true_chunks)){

                     , n_false := NULL][

                     , n_min := NULL][

                     , n := nocell_bccount[match(falseBC,nocell_bccount$XC),n]]

  }else{

    binmap <- data.table()

  }

  if(!is.null(opt$barcodes$barcode_sharing)){

    share_table <- data.table::fread( opt$barcodes$barcode_sharing, header = F, skip = 1)

    if(ncol(share_table) > 2){ #flatten table more if necessary

      share_table <- data.table::melt(share_table, id.vars = "V1")[,variable := NULL]

    }

    setnames(share_table, c("main_bc","shared_bc"))

    share_mode <- data.table::fread( opt$barcodes$barcode_sharing, header = F, nrows = 1)$V1

    share_mode <- as.numeric(unlist(strsplit(gsub(pattern = "#",replacement = "", x = share_mode),"-")))

    if(nrow(binmap)>0){ #first fix the noisy BC assignments so that they dont go into share barcodes

      binmap[, partial_bc := substr(trueBC,start = share_mode[1], stop = share_mode[2])] 

      binmap <- merge(binmap, share_table, all.x = TRUE, by.x = "partial_bc", by.y = "shared_bc")

      substr(binmap[!is.na(main_bc)]$trueBC,start = share_mode[1], stop = share_mode[2]) <- binmap[!is.na(main_bc)]$main_bc #replace to main barcode string

      binmap[,c("partial_bc", "main_bc") := NULL]

    }

    #now check for merging detected barcodes

    bc_detected[, partial_bc := substr(XC,start = share_mode[1], stop = share_mode[2])]

    bc_detected <- merge(bc_detected, share_table, all.x = TRUE, by.x = "partial_bc", by.y = "shared_bc")

    share_map <- bc_detected[!is.na(main_bc)]

    share_map[,trueBC := XC]

    substr(share_map$trueBC, start = share_mode[1], stop = share_mode[2]) <- share_map$main_bc

    setnames(share_map,"XC","falseBC")

    share_map[,c("partial_bc","main_bc","cellindex") := NULL][,hamming := 0]

    share_map <- share_map[,c("falseBC","hamming","trueBC","n"), with = FALSE]

    #messy move the shorter true bc list to the main R script

    bc_detected <- bc_detected[is.na(main_bc)]

    bc_detected[,c("partial_bc","main_bc") := NULL]

    bccount <<- bc_detected

    if(nrow(binmap)>0){

      binmap <- rbind(binmap, share_map)

    }else{

      binmap <- share_map

    }

  }

  print(paste("Found",nrow(binmap),"daughter barcodes that can be binned into",length(unique(binmap[,trueBC])),"parent barcodes."))

  print(paste("Binned barcodes correspond to",sum(binmap[,n]),"reads."))

  return(binmap)

source(paste0(opt$zUMIs_directory,"/barcodeIDFUN.R"))

options(datatable.fread.input.cmd.message=FALSE)

data.table::setDTthreads(threads=opt$num_threads)

if(opt$barcodes$barcode_sharing == ""){

  opt$barcodes$barcode_sharing <- NULL

}

#######################################################################

#######################################################################

fwrite(bccount,file=paste0(opt$out_dir,"/zUMIs_output/",opt$project,"kept_barcodes.txt"))

#check if binning of adjacent barcodes should be run

if(opt$barcodes$BarcodeBinning > 0){

if(opt$barcodes$BarcodeBinning > 0 | !is.null(opt$barcodes$barcode_sharing)){

  binmap <- BCbin(bccount_file = paste0(opt$out_dir,"/", opt$project, ".BCstats.txt"),

                  bc_detected  = bccount)

  fwrite(binmap,file=paste0(opt$out_dir,"/zUMIs_output/",opt$project,".BCbinning.txt"))

                            uiOutput("barcodeUI"),

                            numericInput("HamBC","Hamming distance collapsing of close cell barcode sequences.",value=1,min=0,max=5,step=1),

                            numericInput("nReadsBC","Keep only the cell barcodes with atleast n number of reads",value=100,min=1,max=5,step=1),

                            textInput("sharedBC",label = "Optional: Barcode Sharing (path to file):",value = NULL),

                            checkboxInput("demux", label = "Demultiplex into per-cell bam files?", value = F),

                            shinyBS::bsTooltip(id="demux", title = "Output files will be stored in zUMIs_output/demultiplexed/ .", 

                                               placement = "top", trigger = "hover",options = list(container = "body"))

      "barcodes" = list(

        "barcode_num" = input$BCnum,

        "barcode_file" = input$BCfile,

        "barcode_sharing" = input$sharedBC,

        "automatic" = ifelse(input$barcodeChoice=="Automatic", TRUE, FALSE),

        "BarcodeBinning" = input$HamBC,

        "nReadsperCell" = input$nReadsBC,

      updateNumericInput(session = session, inputId = "HamBC", value = ya$barcodes$BarcodeBinning)

      updateNumericInput(session = session, inputId = "nReadsBC", value = ya$barcodes$nReadsperCell)

      updateTextInput(session = session, inputId = "sharedBC", value = ya$barcodes$barcode_sharing)

      updateCheckboxInput(session = session, inputId = "demux", value = ya$barcodes$demultiplex)

      if(!is.null(ya$read_layout)){

# Pipeline to run UMI-seq analysis from fastq to read count tables.

# Authors: Swati Parekh, Christoph Ziegenhain, Beate Vieth & Ines Hellmann

# Contact: sparekh@age.mpg.de or christoph.ziegenhain@ki.se

vers=2.7.1d

vers=2.7.2

currentv=`curl -s https://raw.githubusercontent.com/sdparekh/zUMIs/master/zUMIs-master.sh | grep '^vers=' | cut -f2 -d "="`

if [ "$currentv" != "$vers" ]; then echo -e "------------- \n\n Good news! A newer version of zUMIs is available at https://github.com/sdparekh/zUMIs \n\n-------------"; fi