Biology¶

Research Computing as the organization that is directly responsible for development and support of the research computing resources on campus - both hardware and software - is fully committed to supporting biological research computing at the University of Florida. As a part of our commitment to this mission we have three Biological Applications Support specialists and a Biological Computing Training specialist.

This section of the HiperGator wiki is dedicated to all things related to biological research computing at UF. We believe that the success of our support of the biological computing efforts at UF depends on the stability of the four main corner stones - hardware resources, software availability and support, training and documentation, and direct consulting and problem solving. The following sections address each of these major points.

Hardware¶

HiperGator nodes are grouped into partitions with different hardware features. The default partition, hpg-default, provides 128 CPUs and 1TB RAM per node which is usually sufficient for most traditional CPU-based computational biology applications. Large memory servers are available in the bigmem queue and provide up to 4TB of RAM. Three GPU partitions are available to serve a rapidly growing list of CUDA-accelerated tools.

The nodeInfo command from the ufrc module can be used to show available node features.

$ nodeInfo
   PARTITION NODES  SOC CPUS MEM(GB) FEATURES
      bigmem     2    8  128    4023 bigmem,amd,rome,infiniband,el9
      bigmem     1    4   96    1509 bigmem,intel,skylake,infiniband,el9
    hpg-b200    30    2  112    2010 ai2,su1,intel,emerald,infiniband,gpu,b200,el9
    hpg-b200    30    2  112    2010 ai2,su2,intel,emerald,infiniband,gpu,b200,el9
 hpg-default   233    8  128    1003 hpg3,amd,rome,infiniband,el9
     hpg-dev     5    8   64     500 hpg3,amd,milan,infiniband,el9
   hpg-milan   145    8   64     500 hpg3,amd,milan,infiniband,el9
   hpg-turin   195    4   96     752 hpg4,amd,turin,infiniband,gpu,l4,el9
       hwgui     3    4   96     752 hpg4,amd,turin,infiniband,gpu,l4,el9

Software¶

Modules¶

To ease the burden of setting up, administering, and maintaining a robust computing environment for the multitude of the applications used by biological researchers we implemented an environmental modules system based on lmod software at UFRC. Over 1000 biology applications are currently available in the module system. Please read documentation on using software via modules or pursue the currently much more extensive documentation written by the lmod author. Users are encouraged to browse the available modules from our list of applications and search for specific tools using in the search box above (top right). Each module is supported by a help page with general information, as well as any relevant additional usage information and/or example job scripts. For an example, please see the ncbi-fcs module.

Software installation¶

If you research requires installation of a particular application or a suite of software whether command-line (CLI) or GUI based we're here to help. Please submit a software installation request in the support system with a link to the source code and we'll promptly review and if possible install and configure the application for you as well as provide a module. We will also host any reference databases required in our dedicated Reference Data Collection.

Software Support¶

We provide support for building your own software on HPC resources as well as for particular job troubleshooting problems encountered in the day-to-day usage of the biocomputing software at UFRC. Please submit a help request in the support system at any time. We also provide data analysis consulting through our team of bioinformatics scientists. They will also provide any software support required for the data analysis pipelines that emerge from consultation.

Databases¶

UFRC maintains a repository of reference data that can be accessed by all HiPerGator users. The primary purposes of this repository are researcher convenience, efficient use of filesystem space, and cost savings. We are happy to download and build reference datasets and configure applications installed on HiPerGator to automatically make use of the available reference data. Having UFRC host common reference data means that a research group does not have to use their Blue or Orange quota to host redundant copies of common reference data.

The UFRC database collection covers several major categories of biological research:

Alignment & Mapping Indices: Pre-built indices for tools like blast, bowtie2, bwa, hisat2, star, and kallisto.
Reference Genomes & Databases: 1000genomes, refseq, silva_rrna, uniprot, swissprot, pfam, dbsnp, clinvar, and pdb70.
Metagenomics & Taxonomic Classifiers: Databases for kraken2, kaiju, metaphlan, humann, and gtdbtk.
Structural & Functional Biology: Specialized resource folders for tools like alphafold, colabfold, rosettafold, and foldseek.
Single-Cell & Long-Read Genomics: References packaged for cellranger, cellranger-arc, and nanopore.