Software

Pre-installed scientific applications on the cluster are managed with Lmod environment modules. Loading a module puts the program on your PATH and pulls in any dependencies it needs (compilers, Java, MPI, CUDA, and so on).

This section documents production science codes only (genomics, molecular dynamics, modeling, quantum chemistry). Compilers, MPI, CUDA, CMake, Conda, Java, and Python are loaded as dependencies; they are not given their own guides. List everything with module avail.

Application pages are alphabetical in the sidebar. Each page covers one module (or a family of versions): what it is, how to load it, a minimal job, and cluster-specific caveats.

Warning

Do not run these tools on the login node. Submit a Slurm job (sbatch or salloc). See Getting Started.

Genomics

BWA

Short-read DNA alignment

FastQC

Quality control of FASTQ files

GATK

Variant discovery toolkit

MultiQC

Aggregate QC reports into one HTML file

samtools

SAM/BAM/CRAM processing

SnpEff

Variant effect annotation

Softepigen

MS-HRM primer design

SRA Toolkit

Download and convert NCBI SRA runs

STAR

Splice-aware RNA-seq aligner

TElocal

Locus-level transposable element quantification

Trimmomatic

Adapter and quality trimming of Illumina reads

Molecular dynamics

Amber

Biomolecular simulation suite

CHARMM

Macromolecular mechanics

Desmond

GPU molecular dynamics (D. E. Shaw)

GROMACS

Molecular dynamics with CUDA

LAMMPS

Classical MD for materials

NAMD

Parallel biomolecular MD

OpenMM

MD toolkit (Python / CUDA)

Molecular modeling

Open Babel

Chemical file format conversion

PLUMED

Enhanced sampling and MD analysis

Schrödinger

Commercial modeling suite

VMD

Visualization and analysis

Quantum chemistry

DFTB+

Density-functional tight binding

JDFTx

Plane-wave DFT

Multiwfn

Wavefunction analysis

ORCA

Quantum chemistry package

Q-Chem

Ab initio quantum chemistry

VASP

Plane-wave DFT (Zen 3 build)

xTB

Semiempirical tight binding

Using modules

List everything that is installed:

$ module avail

Filter by category or name:

$ module avail genom
$ module avail md
$ module avail qm
$ module spider STAR
$ module whatis genom/fastqc
$ module help genom/fastqc

Load and unload:

$ module load genom/fastqc
$ module list
$ module unload genom/fastqc
$ module purge

If you omit the version, Lmod loads the default (marked (D) in module avail). Several packages have variants in the version string (+cuda, +avx2, +zen3). Load the variant you need; the default is not always the GPU build.

Note

Modules apply only to the current shell. Load them again in every new session, and always load them inside the job script so the compute node sees the same environment.

You can load several modules in one command when the tools do not share a Python/conda prefix. MultiQC, TElocal, and OpenMM each have their own prefix; you do not need lang/conda to run them, and you should not mix those prefixes in the same job.

Categories group related packages (genom, md, mm, qm, lang, …). Tab-complete after module load md/ to see what is available.

Storage reminders

Location	Variable	Use for
Home	`$HOME`	Config and small files only
Data	`$DATA_DIR`	Inputs, indices, results you must keep
Scratch	`$SCRATCH_DIR`	Temporary job files (deleted after a week)

Details: Storage.

Example RNA-seq / TE workflow

These tools are often used in separate jobs, not all loaded at once. A typical sequence (adjust to your protocol):

SRA Toolkit — prefetch + fasterq-dump into scratch, keep FASTQ on $DATA_DIR if needed.
FastQC — QC of raw reads.
Trimmomatic — adapters and quality.
FastQC again on trimmed reads (optional).
STAR — build the index once on $DATA_DIR; map with multimappers if you will run TElocal.
TElocal — counts from the BAM (download .locInd yourself).
MultiQC — one HTML report from FastQC/STAR logs.

Each step is a sbatch script on the corresponding page. Do not chain a multi-hour STAR run and a download on the login node.

For DNA short reads, a common chain is BWA → samtools → GATK → SnpEff.

Building software

You may compile tools in your own directories ($HOME or $DATA_DIR) if they are not provided as modules. Keep large builds and test runs off the login node.

Do not install into /opt/ohpc/pub; that tree is reserved for cluster-wide modules.

Request a software

If a package should be available to everyone as a module, email the administrator at fadasme@ucm.cl with:

Name and version (or git tag)
Homepage or repository
Why it is needed and who will use it
Any license constraints