Saturday, October 15, 2011

Fwd: Release of CGA Tools 1.5

Release of CGA Tools 1.5:

Hi everyone,

We are excited to announce the launch of CGA Tools 1.5,
which supports data manipulation and downstream analyses of files produced by
the recently introduced Analysis Pipeline version 2.0.  Some new features
and enhancements for CGA Tools 1.5 include:

  • Addition of varfilter
    (beta) tool to facilitate filtering of variants based on specific
    attributes, such as call confidence and variant type 
  • Enhancement to generatemastervar
    (beta) tool to include additional CNV information provided in Analysis Pipeline
  • Improved somatic variant
    detection and scoring method in calldiff somatic output option
  • Improvements in data
    representation in map2sam and evidence2sam (beta)

NOTE: Some of the tools above are still in beta. We welcome
feedback on the usability of these tools.

The source files and binaries for CGA Tools 1.5 can be
downloaded at  CGA Tools 1.5 Release Notes and CGA Tools
User Guide can be downloaded from SourceForge and can also be found on the User
Community Knowledge Base. 

We hope that you will find these tools useful in helping
determine the variants that matter!  As always, we would love to hear your


Pam Tang

Sr. Product Manager, Analysis Pipeline


new command: varfilter
cgatools varfilter - Copies input var file or masterVar file to output, applying specified filters.


Copies input var file or masterVar file to output, applying specified filters (which are available to all cgatools commands that read a var file or masterVar file as input). Filters are specified by appending the filter
specification to the var file name on the command line.

For example: /path/to/var.tsv.bz2#varQuality!=VQHIGH

The preceding example filters out any calls marked as VQLOW. The filter specification follows the "#" sign, and consists of a list of filters to apply, separated by a comma. Each filter is a colon-separated list of call selectors. Any scored call that passes all the colon-separated call selectors for one or more of the comma-separated filters is turned into a no-call.

The following call selectors are available:

* 'hom Selects only calls in homozygous loci.'
* 'het Selects any scored call not selected by the hom selector.'
* 'varType=XX Selects calls whose varType is XX.'
* 'varScoreVAR* 'varScoreEAF* 'varQuality!=XX Selects calls whose varQuality is not XX.'

Here is an example that filters homozygous SNPs with varScoreVAF < 25 and heterozygous insertions
with varScoreEAF < 50:

(Via CGI user forum.)