bio-ngs¶ ↑

Provides a framework for handling NGS data with Bioruby.

What we want to do and support¶ ↑

• SAMtools

• BWA

• Bowtie/TopHat/Cufflinks

• Reporting: text and graphs

• SGE?

Tasks¶ ↑

We’ll try to keep this list updated but just in case type ‘biongs -T’ to get the most updated list NOTE: We are working on these and other tasks, if you find some bugs, please open an issue on Github.

bwa¶ ↑

• biongs bwa:aln:long [FASTQ] –file-out=FILE_OUT –prefix=PREFIX # Run the aligment for LONG query sequences

• biongs bwa:aln:short [FASTQ] –file-out=FILE_OUT –prefix=PREFIX # Run the aligment for SHORT query sequences

• biongs bwa:index:long [FASTA] # Make the BWT index for a LONG FASTA database

• biongs bwa:index:short [FASTA] # Make the BWT index for a SHORT FASTA database

• biongs bwa:sam:paired –fastq=one two three –file-out=FILE_OUT –prefix=PREFIX –sai=one two three # Convert SAI alignment output into SAM format (paired ends)

• biongs bwa:sam:single [SAI] –fastq=FASTQ –file-out=FILE_OUT –prefix=PREFIX # Convert SAI alignment output into SAM format (single end)

convert¶ ↑

Most of this tasks create sub-processes to speed up conversions

• biongs convert:bam:extract_genes BAM GENES –ensembl-release=N # Extract GENES from bam. It connects to Ensembl Humnan, release 61

• biongs convert:bam:sort BAM [PREFIX]

• biongs convert:bcl:qseq:convert RUN OUTPUT [JOBS] # Convert a bcl dataset in qseq

• biongs convert:illumina:fastq:trim_b FASTQ # perform a trim on all the sequences on B qualities with Illumina’s criteria. Ref to CASAVA manual.

• biongs convert:qseq:fastq:by_file FIRST OUTPUT # Convert a qseq file into fastq

• biongs convert:qseq:fastq:by_lane LANE OUTPUT # Convert all the file in the current and descendant directories belonging to the specified lane in fastq. This command is specific for Illum…

• biongs convert:qseq:fastq:by_lane_index LANE INDEX OUTPUT # Convert the qseq from a line and index in a fastq file

• biongs convert:qseq:fastq:samples_by_lane SAMPLES LANE OUTPUT # Convert the qseqs for each sample in a specific lane. SAMPLES is an array of index codes separated by commas lane is an integer

filter¶ ↑

• biongs filter:by_list TABLE LIST # Extract from TABLE rows with a key in LIST

quality¶ ↑

• biongs quality:boxplot FASTQ_QUALITY_STATS # plot reads quality as boxplot

• biongs quality:fastq_stats FASTQ # Reports quality of FASTQ file

• biongs quality:illumina_b_profile_raw FASTQ –read-length=N # perform a profile for reads coming fom Illumina 1.5+ and write the report in a txt file

• biongs quality:illumina_b_profile_svg FASTQ –read-length=N # perform a profile for reads coming fom Illumina 1.5+

• biongs quality:illumina_project_stats # Reports quality of FASTQ files in an Illumina project directory

• biongs quality:reads FASTQ # perform quality check for NGS reads

• biongs quality:reads_coverage FASTQ_QUALITY_STATS # plot reads coverage in bases

• biongs quality:trim FASTQ # trim all the sequences

rna¶ ↑

• biongs rna:idx2fasta INDEX FASTA # Create a fasta file from an indexed genome, using bowtie-inspect

• biongs rna:tophat DIST INDEX OUTPUTDIR FASTQS # run tophat as from command line, default 6 processors

root¶ ↑

• biongs project NAME

sff¶ ↑

• biongs sff:extract [FILE] # Run sff_extract on a SFF file

TasksExamples¶ ↑

Conversion¶ ↑

biongs convert:bam:extract_genes your_original.bam BLID,GATA3,PTPRC --ensembl_release=61 --ensembl_specie=homo_sapiens

Filtering¶ ↑

When you have your mapped reads to a reference genome, you can decide to filter the output (GTF) to extract only those transcripts which have your desired requirements. You can filter for lenght, if it’s multi or mono exon, the coverage, if it’s a brand new transcript or an altrady annotated gene but with a new isoform or just the annotated transcripts.

Scenario: filtering transcripts
Having a transcripts.gtf dataset generated from CufflinksQuantification
I want a only the new transcripts (also with an annotated gene)
Which are multi exons
With a lenght greater than 1340
With minimum coverage greater than 10
Then I want to save them in my_filtered_data.gtf

biongs filter:cufflinks:transcripts your_original.gtf -m -l 1340 -c 10.0 -n -o my_filtered_data.gtf

Then in some case I need to extract only some of them or maybe parsing them from external programs. Biongs has a specific trask for this:

Having my_filtered_data.gtf
Generated by "filtering transcripts"
I want to extract transcript number 10
Then I want to save it in BED format
Using UCSC notation

biongs filter:cufflinks:tra_at_idx my_filtered_data.gtf #of_the_transcript_to_retrieve -u

The first time tra_at_idx is used, it will take more time than usual becase it creates an internal index: a simple HASH mashalled and dumped, stored in a file with the name similar to the imput with an idx as postfix.

ForDevelopers¶ ↑

Contribute ¶ ↑

Clone Main Repository¶ ↑

This command will crate a local copy of the main repository

git clone https://github.com/helios/bioruby-ngs

Install Bioinformatics Tools into the repository directory¶ ↑

rake devenv:bio_tools

Wrapper¶ ↑

Bio-Ngs comes with a build-in wrapper to map binary software directly in BioRuby as objects. From this wrapper object is possible to create Thor task as well, with a lot of sugar.

Wrapping a binary¶ ↑

We want wrap TopHat the famous tool for NGS analyses.

* The first step is to include the Wrapping module
* set the name of the binary to call. Note: if you avid to set the program name it would not be possible to create a thor task and/or run the program
* add the options that the binary accepts, usually if preferred to declare all the options, discover them typing 'your_program_name -h'

   module Bio
     module Ngs    
       class Tophat
         include Bio::Command::Wrapper

         set_program Bio::Ngs::Utils.binary("tophat/tophat")
         add_option "output-dir",:type => :string, :aliases => '-o'
         add_option "min-anchor", :type => :numeric, :aliases => '-a'
         add_option "splice-mismatches", :type => :numeric, :aliases => '-m'
         #all other options that you want to expose with the wrapping
       end #Tophat
     end #Ngs
   end #Bio

is possible to use specify in the class

use_aliases

if you want to give a priority to short notation or if your program has only the short notation but you want to extend the task with the long one as well. We defined a new property for add_option called

:collapse => true

is used only with use_aliases and it collapse the passed parameter to the short notation - example coming from fastx.rb wrapper, note last row - : module Bio module Ngs

module Fastx class Trim include Bio::Command::Wrapper set_program Bio::Ngs::Utils.binary(“fastq_quality_trimmer”) use_aliases add_option :min_size, :type=>:numeric, :default=>20, :aliases => “-l”, :desc=>“Minimum length - sequences shorter than this (after trimming) will be discarded. Default = 0 = no minimum length.” add_option :min_quality, :type=>:numeric, :default=>10, :aliases => “-t”, :desc=>“Quality threshold - nucleotides with lower quality will be trimmed (from the end of the sequence).” add_option :output, :type=>:string, :aliases => “-o”, :desc => “FASTQ output file.”, :collapse=>true add_option :input, :type=>:string, :aliases => “-i”, :desc => “FASTQ input file.”, :collapse=>true add_option :gzip, :type => :boolean, :aliases => “-z”, :desc => “Compress output with GZIP.” add_option :verbose, :type => :boolean, :aliases => “-v”, :desc => “[-v] = Verbose - report number of sequences. If [-o] is specified, report will be printed to STDOUT. If [-o] is not specified (and output goes to STDOUT), report will be printed to STDERR.” add_option :quality_type, :type=>:numeric, :default => 33, :aliases => “-Q”, :desc=>“Quality of fastq file” end end end end fastq_quality_trimmer accepts only short notation options and we need to pass an input file, but for some reason popen used internally doesn’t work properly with the standard behavior so using :collapse=>true the application will be called:

fastq_quality_trimmer -t 20 -t 10 -Q 33 -iinput_file_name.fastq -ooutput_file_name.fastq_trim

running the program by hand form the command line using a space as separator after -i and -o works as expected. :collapse is a work around for this problem.

In case you program work like git which has a main program and the sub_programs for each feature you can use specify the sub program name with

set_sub_program "sub_name"

The wrapper will run the command composing:

set_program set_sub_program options arguments

A practical example of this behavior is samtools which has multiple sub programs view, merge, sort, .… SamTools is a particular case because in biongs we are using bio-samtools a binding with FFI and the wrapper because the merge function was too complicated for the binding or at least we do not spent enough time on it, so we make the wrapping for this functionality.

This step is very similar to define a Thor task, add_option is grabbed/inspired from Thor. Then you can user this binary also from a bioruby script just calling: tophat = Bio::Ngs::Tophat.new tophat.params = {“mate-inner-dist”=>dist, “output-dir”=>outputdir, “num-threads”=>1, “solexa1.3-quals”=>true} #very important: you can pass parameters that have a name which has been previously declared in the Tophat’s class. # if you want to pass not declared parameters/options please use arguments. tophat.run :arguments=>[index, “#{fastqs}” ]

Define the Task¶ ↑

With our new wrapper, let’s define a Thor task on the fly

class MyTasks < Thor
  desc "tophat DIST INDEX OUTPUTDIR FASTQS", "run tophat as from command line, default 6 processors"
  Bio::Ngs::Tophat.new.thor_task(self, :tophat) do |wrapper, task, dist, index, outputdir, fastqs|
    wrapper.params = {"mate-inner-dist"=>dist, "output-dir"=>outputdir, "num-threads"=>1, "solexa1.3-quals"=>true}
    wrapper.run :arguments=>[index, "#{fastqs}" ], :separator=>"="
    #you tasks here
  end
end

Now is you list the tasks with ‘thor -T’ you will see the new task.

You can create a new wrapper and configure it and run it from inside a Thor’s tasks, like in ‘biongs quality:boxplot’

desc "boxplot FASTQ_QUALITY_STATS", "plot reads quality as boxplot"
method_option :title, :type=>:string, :aliases =>"-t", :desc  => "Title (usually the solexa file name) - will be plotted on the graph."
method_option :output, :type=>:string, :aliases =>"-o", :desc => "Output file name. default is input file_name with .txt."
def boxplot(fastq_quality_stats)
  output_file = options.output || "#{fastq_quality_stats}.png"
  boxplot = Bio::Ngs::Fastx::ReadsBoxPlot.new
  boxplot.params={input:fastq_quality_stats, output:output_file}
  boxplot.run
end

Override the run command when the binary dosen’t behave normally¶ ↑

module Bio
  module Ngs    
    module Samtools
      class View
        include Bio::Command::Wrapper
        set_program Bio::Ngs::Utils.binary("samtools")
        add_option "output", :type => :string, :aliases => '-o'

        alias :original_run :run
        def run(opts = {:options=>{}, :arguments=>[], :output_file=>nil, :separator=>"="})
          opts[:arguments].insert(0,"view")
          opts[:arguments].insert(1,"-b")
          opts[:arguments].insert(2,"-o")
          original_run(opts)
        end
      end #View
    end #Samtools
  end #Ngs
end #Bio

Disable binary check at load time¶ ↑

When a wrapping is defined BioNGS verify that the program is installed on the local system, if it is not it thrown an warning message and the task is disabled by default. This check is made for each binary wrapped, so it could takes long the first time you load BioNGS. To skip this check the user can define an environment variable assigning one of these terms “true yes ok 1” to BIONGS_SKIP_CHECK_BINARIES

export BIONGS_SKIP_CHECK_BINARIES=true

you can also add this setting to the .bashrc or .profile in the user home directory.

Features¶ ↑

Iterators for output files¶ ↑

Example CuffDiff. In this class is possible to define an iterator for a specific set of output files: genes, isoforms, tss_groups, cds. To activate the iterator is just a matter of call a class method in the class definition

class Bio::Ngs::Cufflinks::Diff
  #... all the previous definitions
  #define iterators
  add_iterator_for :genes
  add_iterator_for :isoforms
  add_iterator_for :cds
  add_iterator_for :tss_groups
end

This is an example of CuffDiff, parsing genes.fpkm_tracking file:

Bio::Ngs::Cufflinks::Diff.foreach_gene_tracked("path_to_cuffdiff_output_directory") do |gene_fpkm_track|
  expression_profile = (1..7).map do |sample_idx|
    gene_fpkm_track["q#{sample_idx}_FPKM"].to_f
  end

  #do your stuff accessing this tabular file with gene_fpkm_track["name of the field"]
end

In this case internally CSV library has been used to parse in an easy way the file, there is a lack of performances with huge files, gaining in flexibility.

Loading or Not tasks from outside¶ ↑

If in your external library or binary you define LoadBaseTasks in Bio::Ngs (as a costant) requiring ‘bio-ngs’ bio-ngs’s tasks will not load but only the libraries.

module Bio
  module Ngs
    LoadBaseTasks = true
  end
end

This is something useful if you want to develop a separate binary which uses bio-ngs librariys. Is not yet possible to define a list of desired tasks to load.

Notes¶ ↑

• It’s possible to add more sugar and we are working hard on it

• aliases are not well supported at this time. ToDo

REQUIREMENTS¶ ↑

• hannonlab.cshl.edu/fastx_toolkit/ (the gem tries to install this tool by itself)

• www.gnuplot.info/ tested on version 4.6

• libxslt1-dev

Pleas follow the instruction for your own distribution/operating system

TODO¶ ↑

• Write Tutorial for Wrapper & Pipes

• Write Tutorial for handling Illumina/Fastq.gz with BioNGS Bio::Ngs::Illumina::FastqGz

• Report the version of every software installed/used from bio-ngs

• Develop fastq quality reports with RibuVis ?

• Write documentation

• DONE: Wrapper: better support for aliases and Wrapper#params

• Convert: re factor code to use ::Daemons

• DONE:misk_tasks? Extract genes/regions of interest from a bam file and create a smaller bam

• BRANCH:misk_tasks Explore possibility to user DelayedJobs

• biongs ann:ensembl:gtf:features:categorize GTF GTF categorize also by chromosome not only by BioType

• configuration file input,output, experimental design

• DONE: include fastx toolkit, download and compile

• ANSWER: how to put in background tasks that can be run in parallel? Use Parallel (see code for quality:illumina_project_stats)

• is it possible to establish a relation between input data and output data ? like fastq task_selected output/s

• add description for developers on howto include news external tool with versions.yaml

ChangeLog¶ ↑

* 2011.05-26: Bump to version 0.2.0 Complete support for installing fastx and possibly other downloadable tool, inside the gem 
* 2011-05-25: Bump to version 0.1.0 Update Cufflinks toolkit 1.0.2. Added initial support to fastx tool kit (binaries not included)
* 2011-04-08: Tasks for filtering Ensembl annotation and create classifications. (misk_tasks branch)

Contributing to bio-ngs¶ ↑

Please do not hesitate to contact us:

Raoul J.P. Bonnal, github.com/helios, r -at- bioruby -dot- org Francesco Strozzi, github.com/fstrozzi

• Check out the latest master to make sure the feature hasn’t been implemented or the bug hasn’t been fixed yet

• Check out the issue tracker to make sure someone already hasn’t requested it and/or contributed it

• Fork the project

• Start a feature/bugfix branch

• Commit and push until you are happy with your contribution

• Make sure to add tests for it. This is important so I don’t break it in a future version unintentionally.

• Please try not to mess with the Rakefile, version, or history. If you want to have your own version, or is otherwise necessary, that is fine, but please isolate to its own commit so I can cherry-pick around it.

Copyright¶ ↑

Copyright © 2011 Francesco Strozzi and Raoul J.P. Bonnal. See LICENSE.txt for further details.