#

Application Types

This page summarizes the application types available on the FAS RC Cluster. Individual applications can be searched via
the RC Portal search tool

HeLmod CentOS 7

  This is the default Lmod-based module system on the FASRC cluster. Tools in this application type were built using the standard cluster CentOS 7 operating system as a base. This is enabled for every cluster user so no additional activation steps are required. More details can be found on the RC Website HeLmod page.

HeLmod applications are activated via Lmod-style modules that may either be a single module name:

module load zlib/1.2.11-fasrc01

Or by a series of modules that enable compiler or MPI specific branches

module load gcc/8.2.0-fasrc01 zlib/1.2.11-fasrc01

 

Easy Build

 

The Easy Build system is a third party Lmod-based application build and deployment system, described in detail both in general and for the FASRC cluster specifically. Easy Build modules are available by default on the FASRC cluster so no additional activation is required.

Easy Build applications are activated via Lmod-style modules like HeLmod but are constructed from pre-made recipes. A module load using the "intel" toolchain might look like this:

module load intel/2017c impi/2017.4.239 SAMtools/1.9

 

x86_84

  x86_64 applications are simple, pre-compiled Linux binaries. These may either be downloaded and used directly, or maybe part of a *.tar.gz file that must be unpacked. Like any Linux binary, these applications can be used either by specifying the file along with it's path, or by placing the directory containing the application in your PATH environment variable.

For example

$ wget https://data.broadinstitute.org/alkesgroup/BOLT-LMM/downloads/BOLT-LMM_v2.3.2.tar.gz
$ tar xvf BOLT-LMM_v2.3.2.tar.gz
$ export PATH=`pwd`/BOLT-LMM_v2.3.2:$PATH
$ bolt -h
+-----------------------------+
|                       ___   |
|   BOLT-LMM, v2.3.2   /_ /   |
|   March 10, 2018      /_/   |
|   Po-Ru Loh            //   |
|                        /    |
+-----------------------------+

Copyright (C) 2014-2018 Harvard University.
Distributed under the GNU GPLv3 open source license.

Compiled with USE_SSE: fast aligned memory access
Compiled with USE_MKL: Intel Math Kernel Library linear algebra
Boost version: 1_58

Command line options:

./bolt -h

 

Anaconda

 

Anaconda is a Python distribution that includes a combination of virtual environment tools and a package management system. The default python and python3 modules on the FASRC cluster are based on Anaconda and the use is described elsewhere in detail. In general, Anaconda is the default, recommended solution for Python packages, particularly scientific packages like NumPy, SciPy, and pandas, and can increasingly be used for R environments as well.

Once you have created an Anaconda environment and activated it, package installation is usually straightforward.
For example

(fasrc)$ conda install pandas==0.25.1

 

Will install the pandas package of the specified version

Bioconda

 

Bioconda is an Anaconda channel devoted to packages for bioinformatics tools. In addition to things like SAM Tools and GATK, an entire suite of R and R packages
are available.

Activation of Bioconda packages requires the creation of an Anaconda environment. Once this is done, several Anaconda "channels" must be configured to allow installation from the Bioconda repository.

 
# If you don't have a local conda environment...
module load python
conda create -n fasrc pip
source activate fasrc
# Configure the bioconda channels
conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge

 

Open OnDemand

 

While graphical applications can be run by X11 port-forwarding, the Open OnDemand platform is better suited to interactive applications like Matlab and RStudio or just plain Linux desktop. These applications can be accessed via vdi.rc.fas.harvard.edu. Open OnDemand replaces the previously supported NoMachine services.

 

Java

 

Some applications are distributed as Java Jar files. Jar files can usually be run directly with an appropriate Java interpreter. The system java on FAS RC resources is often too old to run recently developed applications, so a JDK module must often be loaded.

For example, the Broad Institute Picard Tools can be downloaded and run directly with a Java 1.8 module loaded.

$ module load jdk/1.8.0_172-fasrc01
$ wget https://github.com/broadinstitute/picard/releases/download/2.20.8/picard.jar
$ java -jar picard.jar -h
(base) FAS-AARON:temp akitzmiller$ java -jar picard.jar -h
USAGE: PicardCommandLine  [-h]

Available Programs:
                

CC BY-NC-SA 4.0 This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Permissions beyond the scope of this license may be available at Attribution.