NAME
Tutorial_Pipeline02.pl - Another example pipeline for the ViennaNGS toolbox
SYNOPSIS
perl Tutorial_Pipeline02.plDESCRIPTION
This script is a showcase for using Bio::ViennaNGS components with a real-world NGS example.
We start from a file containing ENSEMBL annotation information for human protein-coding genes, which have a read pileup of at least 1001 reads in an ENCODE dataset mapped with segemehl. We are insterested in visualizing those genes together with a 50nt region upstream of the gene start. As such, will go through the individual steps required for preparation of the files and subsequent UCSC Track Hub visualization.
PREREQUITES
For running this tutorial on your machine you will need a full installation of the Bio::ViennaNGS distribution, including all third party dependencies, i.e. a recent version of bedtools as well as the bedGraphToBigWig utility from the UCSC source distribution (available here). In addition, the following input files (which can be downloaded here) will be used throughout this tutorial:
DISCLAIMER
This tutorial works on a real-world biological data set of several gigabytes in size i.e. the analysis will eventually take a few hours to finish, depending on your hardware. If you run this script locally you need to ensure that your system has enough hardware resources available.
PIPELINE
Let's first initialize some variables and generate a chromosome_sizes hash.
my $bed = 'hg19_highlyexpressed.bed';
my $name = (split(/\./,$bed))[0];
my $upstream = 50;
my $outfilebn = $name.".ext".$upstream."_upstream";
my $outfilebed = $outfilebn.".bed";
my %sizes = %{fetch_chrom_sizes('hg19')};Generate a Bio::ViennaNGS::FeatureChain object
We'll now parse our BED file of interest, generate a feature array and pass it on to Bio::ViennaNGS::FeatureChain, thereby creating a new FeatureChain Moose object that contains the original BED entries.
my @featurelist = @{parse_bed6($bed)};
my $chain = Bio::ViennaNGS::FeatureChain->new('type'=>'original','chain'=>\@featurelist);Extend the existing chain for UCSC visualization
The FeatureChain object will now be extended 50nt upstream of the gene start to retrieve a BED file containing both the genic and potential promoter regions.
my $extended_chain = extend_chain(\%sizes,$chain,$upstream,0,0,0);We'll also extend the entire U6 gene span 50nt upstream for later usage.
my $extended_chain2 = extend_chain(\%sizes,$chain,$upstream,0,0,0);Print extended Bio::ViennaNGS::FeatureChain objects to files
Extended chains are now printed out to make them available for external tools like bedtools.
open (my $Out, ">",$outfilebed) or die "$!";
my $out = $extended_chain->print_chain();
print $Out $out;
close($Out);Summary of so far used methods
fetch_chrom_sizes()-
Returns a chromosome-sizes hash reference for the specified species, e.g. hg19, mm9, mm10, etc.
parse_bed6()-
Reads a bed6 file and returns a feature array.
Bio::ViennaNGS::FeatureChain->new()-
Generates a new Bio::ViennaNGS::FeatureChain object from a feature array
extend_chain()-
Extends a Bio::ViennaNGS::FeatureChain object by given constraints
Sequence processing and analysis
We will now generate FASTA files from the extended BED files by using the bedtools getfasta method.
To analyze putative sequence motifs in the newly generated Fasta files, we analyze the k-mer content using the Bio::ViennaNGS kmer_enrichment() method for k-mers of length 6 to 8 nt.
my $hg19fa = "hg19_chromchecked.fa";
my $outfilefa = $outfilebn.".fa";
my $bedtools = `bedtools getfasta -s -fi $hg19fa -bed $outfilebed -fo $outfilefa`;
print STDERR "$bedtools\n" if $?;
open(IN,"<", $outfilefa) || die ("Could not open $outfilefa!\n@!\n");
my @fastaseqs;
while(<IN>){
chomp (my $raw = $_);
next if ($_ =~ /^>/);
push @fastaseqs, $raw;
}
close(IN);
for (6..8){
my %kmer = %{kmer_enrichment(\@fastaseqs, $_)};
my $total = sum values %kmer;
### Print Output
open(KMER,">","$_\_mers") or die "Could not open file $_\_mers$!\n";
print KMER "$_\-mer\tCount\tRatio\n";
print KMER "TOTAL\t$total\t1\n";
foreach my $key (sort {$kmer{$b} <=> $kmer{$a} } keys %kmer) {
my $ratio = nearest(.0001,$kmer{$key}/$total);
print KMER "$key\t$kmer{$key}\t$ratio\n";
}
close(KMER);
}Retrieve mapped sequences that overlap the extended chain as BAM using bedtools
To generate a BAM file containing uniquely mapped reads that overlap our genes of interest, we use the bedtools intersect method.
my $bam_upstream = $outfilebn."_overlapping_C1R1.bam";
my $cmd = "intersectBed -s -split -abam C1R1.bam -b $outfilebed > $bam_upstream ";
my( $success, $error_message, $full_buf, $stdout_buf, $stderr_buf ) = run(command => $cmd, verbose => 0);
if(!$success){
print STDERR "ERROR: Bedtools intersect unsuccessful\n";
print join "", @$full_buf;
}Split BAM by strand
We will now split the BAM file by strands using the bam_split() routine, thereby creating two new BAM files: One containg all reads mapped to the [+] strand, and one that contains all reads that map to the [-] strand.
my $reversed = 1;
my $wantuniq = 0;
my $wantbed = 1;
my $outdir = cwd();
my $lf = "log.txt";
my @result = split_bam($bam_upstream,$reversed,$wantuniq,$wantbed,$outdir,$lf);
my $bam_p = $result[0]; # BAM file containing fragments of [+] strand
my $bam_n = $result[1]; # BAM file containing fragments of [-] strand
my $size_p = $result[2]; # of alignments on [+] strand
my $size_n = $result[3]; # of alignments on [-] srand
my $bed_p = $result[4]; # BED file containing fragments of [+] strand
my $bed_n = $result[5]; # BED file containing fragments of [-] strandbam_split returns an array, (@ref in our example) that contains six fields: The paths of the BAM files for [+] and [-] strand, number of alignments in the [+] and [-] BAM files as well as the paths to the interim BED files for [=] and [-] strand, respectively.
Create BigWig coverage profiles
Now that we have separate BAM files for each strand at hand, the next step is to create coverage profiles in BigWig format for subsequent UCSC visualization. The routine of choice for this task is bed_or_bam2bw(), which is called separately for [+] and [-] strand.
my $od = cwd();
my $cs_in = "hg19.chrom.sizes";
my $wantnorm = 0;
my $scale = 10000000;
bed_or_bam2bw("bed",$bed_p,$cs_in,"+",$od,$wantnorm,$size_p,$scale,$lf);
bed_or_bam2bw("bed",$bed_n,$cs_in,"-",$od,$wantnorm,$size_n,$scale,$lf);Once finished, there will be two new BigWig files in the current working directoy: hg19_highlyexpressed.pos.bw and hg19_highlyexpressed.neg.bw contain coverage information of the [+] and [-] strand, respectively and will be used for genome browser visualization in the following sections. Besides BigWig files, bed_or_bam2bw() creates BED and BAM files for each strand. We will, however not use these files throughout this tutorial.