Scientific Supercomputing at the NIH

Beagle on Helix

BEAGLE is a software program for imputing genotypes, inferring haplotype phase, and performing genetic association analysis. BEAGLE is designed to analyze large-scale data sets with hundreds of thousands of markers genotyped on thousands of samples. BEAGLE can

Beagle was developed by Brian Browning at the University of Auckland, NZ. Beagle website.

Using Beagle on Helix

Beagle is a Java program. It is important to use the correct version of Java as in the example below.

Sample session (user input in bold). This run uses the sample data in /usr/local/beagle/example/imputation/

helix% /usr/local/jdk/bin/java -Xmx1000m -jar /usr/local/beagle/beagle.jar markers=hapmap.markers phased=hapmap.phased.bgl unphased=hapmap.unphased.bgl missing=? log=imputation
Beagle version 3.0.2

Command line: java -jar beagle.jar 
  markers=hapmap.markers
  phased=hapmap.phased.bgl
  unphased=hapmap.unphased.bgl
  missing=?
  log=imputation

number of markers    =     100
number of haplotypes =       4 (hapmap.unphased.bgl)
number of haplotypes =     116 (hapmap.phased.bgl)

Phasing: iteration 1
Phasing: iteration 2
Phasing: iteration 3
Phasing: iteration 4
Phasing: iteration 5
Phasing: iteration 6
Phasing: iteration 7
Phasing: iteration 8
Phasing: iteration 9
Phasing: iteration 10

Running time for phasing: 1 seconds
Beagle version 3.0.2 finished

This example runs Beagle using the sample data, and limits it to 1000 Mb. Each user is limited to 32GB of memory on Helix.

If you expect to use Beagle frequently, it will be convenient to set up an alias as below:

Bash users: add this to your /home/user/.bashrc file

alias beagle='/usr/local/jdk/bin/java -jar /usr/local/beagle/beagle.jar'

tcsh/csh users: add this to your /home/user/.cshrc file

alias beagle '/usr/local/jdk/bin/java -jar /usr/local/beagle/beagle.jar'

You will then be able to run Beagle by typing a command like:

helix% beagle -Xmx5000m markers=hapmap.markers phased=hapmap.phased.bgl unphased=hapmap.unphased.bgl missing=? log=imputation
to run Beagle using 5GB of memory.

Documentation

Beagle documentation (PDF)