FusionSeq Requirements

From GersteinInfo

Revision as of 00:39, 21 August 2010 by Asboner (Talk | contribs)
Jump to: navigation, search
FusionSeq main web page
User documentation main

Contents

Software Requirements

FusionSeq requires several additional packages to be installed in order to carry out the analysis and visualize the results. Moreover, since its modularity, different programs would need specific libraries. Moreover, some data sets are also required for the analysis (see Data Requirements). Here we describe the complete set of tools that one would need to run the analysis as we do in our lab. The modules should be installed in the listed order.

Alignment tools

Please make sure that blat and bowtie executables are part of the PATH, i.e. they can be accessed and executed from any location on your file system. Moreover, make sure that twoBitToFa is also downloaded from the blat package and part of the PATH.

Scientific and bioinformatics libraries

  • GNU Scientific Library (GSL): this library is a required for the compilation of the BIOS.
  • BIOS: this library can be downloaded as part of RSEQtools, a computational framework to analyze RNA-Seq data, or it can be downloaded as a separate component from here.

Instructions to install GSL and BIOS can be found in Installation_and_Configuration_of_FusionSeq. However, please ensure that you read all the requirements (including Data requirements) and downloaded all the libraries and packages needed.

Drawing tools

  • GD library: The gd library is used to create schematic images of the PE reads connecting the two genes. It is required by gfr2images, which is an optional component of FusionSeq.

Data analysis

  • ROOT: this is a very powerful mathematical and computational framework. In the context of FusionSeq, it is used to perform a Kolomogorov-Smirnov analysis for filtering the breakpoint junctions and plotting the insert-size distribution.
Top

Data Requirements

Here is the list of required data for a comprehensive use of FusionSeq tools.

External

The human genome needs to be properly indexed to be used by bowtie. Please see the instruction of bowtie for performing this operation. Indicatevely, you would need to run something like:

$ bowtie-build -f hg18_nh.fa /path2bowtieIndex/hg18_nh/

where hg18_nh.fa corresponds to the concatenation of all human chromosomes from chromFa.zip without the different haplotypes and "random" stuff.


Provided

The following data sets, bundled in a tarball, can be downloaded here.

  • knownGeneAnnotationTranscriptCompositeModel.txt - the interval file with the coordinates of the composite models
  • knownGeneAnnotationTranscriptCompositeModel.fa - the sequence of all the composite transcripts
  • kgXref.txt - the mapping between the UCSC knownGene annotation set and other information (RefSeq, gene symbols and description etc.)
  • knownToTreefam.txt - the mapping between UCSC knownGene annotation and TreeFam
  • hg18_repeatMasker.interval - the interval file, i.e. the file with the coordinates, of the repetitive regions
  • ribosomal.2bit - the ribosomal sequences in 2bit format

The composite model needs to be indexed by bowtie:

$ bowtie-build -f knownGeneAnnotationTranscriptCompositeModel.fa /path2bowtieIndex/hg18_knownGeneAnnotationTranscriptCompositeModel/hg18_knownGeneAnnotationTranscriptCompositeModel

Although we extensively used the UCSC knownGene annotation set, it is worth mentioning that it is possible to use other gene annotation sets. However, in this case, the same information, and in the same format, should be provided to the corresponding programs.

Personal tools