view tools/spades_3_1_1/plot_spades_stats.xml @ 10:7d72faef9af0 draft

Uploaded
author takadonet
date Fri, 10 Oct 2014 14:40:43 -0400
parents
children
line wrap: on
line source

<tool id="plot_spades_stats" name="SPAdes stats" version="0.1">
  <description>coverage vs. length plot</description>
  <requirements>
    <requirement type="package">R</requirement>  
  </requirements>
  <command interpreter="bash">r_wrapper.sh $script_file</command>

  <inputs>
    <param name="input_scaffolds" type="data" format="tabular" label="Scaffold stats"/>
    <param name="input_contigs" type="data" format="tabular" label="Contig stats"/>
    <param name="length_co" type="integer" value="1000" min="0" label="Length cut-off" help="Contigs with length under that value are shown in red"/>
    <param name="coverage_co" type="integer" value="10" min="0" label="Coverage cut-off" help="Contigs with length under that value are shown in red"/>
  </inputs>
  <configfiles>
    <configfile name="script_file">
## Setup R error handling to go to stderr
options( show.error.messages=F, 
  error = function () { 
    cat( geterrmessage(), file=stderr() ); q( "no", 1, F ) 
} )
files = c("${input_contigs}", "${input_scaffolds}")
types = c("Contigs", "Scaffolds")

## Start plotting device
png("${out_file}", w=500, h=1000)
par(mfrow=c(2,1))

## Loop over the two files
for (i in 1:length(types)){
  seqs = read.table(files[i], header=FALSE, comment.char="#")
  colnames = c("name", "length", "coverage")
  names(seqs) = colnames

  ## Stats over all sequences 
  sl_all = sort(seqs\$length, decreasing=TRUE)
  cs_all = cumsum(sl_all)
  s_all = sum(seqs\$length)
  n50_idx_all = which.min(sl_all[cs_all &lt; 0.5*s_all])
  n90_idx_all = which.min(sl_all[cs_all &lt; 0.9*s_all])
  n50_all = sl_all[n50_idx_all]
  n90_all = sl_all[n90_idx_all]
  
  ## Filter short seqs, redo stats
  seqs_filt = seqs[seqs\$length >= ${length_co} &amp; seqs\$coverage >= ${coverage_co},]
  if (nrow(seqs_filt) > 0){
    sl_filt = sort(seqs_filt\$length, decreasing=TRUE)
    cs_filt = cumsum(sl_filt)
    s_filt = sum(seqs_filt\$length)
    n50_idx_filt = which.min(sl_filt[cs_filt &lt; 0.5*s_filt])
    n90_idx_filt = which.min(sl_filt[cs_filt &lt; 0.9*s_filt])
    n50_filt = sl_filt[n50_idx_filt]
    n90_filt = sl_filt[n90_idx_filt]
  }
  seqs_bad = seqs[seqs\$length &lt; ${length_co} | seqs\$coverage &lt; ${coverage_co},]

  ## Length vs coverage
  plot(length~coverage, data=seqs, log="xy", type="n", main=paste(types[i], ": coverage vs. length", sep=""), xlab="Coverage", ylab="Length")
  if (nrow(seqs_bad) > 0){
    points(length~coverage, data=seqs_bad, cex=0.5, col="red")
  }
  if (nrow(seqs_filt) > 0){
    points(length~coverage, data=seqs_filt, cex=0.5, col="black")
  }
  abline(v=${coverage_co}, h=${length_co}, lty=2, col=grey(0.3))
  legend(x="topleft", legend=c("Before/after filtering", paste(c("N50: ", "N90: ", "Median cov.: "), c(n50_all, n90_all, round(median(seqs\$coverage))), rep("/", 3), c(n50_filt, n90_filt, round(median(seqs_filt\$coverage))), sep="")), cex=0.8)
}
dev.off()
    </configfile>
  </configfiles>
  <outputs>
    <data format="png" name="out_file" />
  </outputs>
  <help>
**What it does**

Using the output of SPAdes (a pair of fasta file and stat file for each of the contigs and scaffolds), it produces a coverage vs. contig plot. Each dot represent a contig/scaffold. Given a coverage and a length cutoff, sequences that do not meet those criteria are shown in red. Some statistics are also given (N50, N90, median contig/scaffold length) both before and after filtering.

Use the "filter SPAdes output" tool to actually filter sequences.
  </help>
</tool>