DBS: RISU: Available Software
RISU provides a variety of computational platforms for faculty and staff to perform their research activities. Ranging from interactive Linux systems used for primary statistical and genetic analysis and high performance compute clusters used for distributed parallel processing to Windows-based servers used for database and statistical processing, DPHS provides a host of resources to accomplish the research goals of its investigators.
Core Interactive Systems:
- (2) Linux Servers - IBM BladeCenter nodes with 8 Dual Core Xeon 2.40 GHz CPUs, and 98 GB RAM, each
High Performance Compute Clusters (HPCC)
The challenge of large scale data processing is addressed by the high-performance computation facilities including:
- 23 Node / 92 core IBM Blade cluster, 828GB RAM
- Access to 140+ Node / 482 core IBM Blade cluster - "Deac Cluster", with a combined 2.8TB RAM.
- DEMON-Isilon (1,358 2.6GHz Intel CPU cores, 4,992 706MHz Nvidia GPU cores, and 190 TB parallel storage space) at WFUHS
- Data Lake - a highly optimized computational resource that includes integrated high performance storage, fast computational resources and optimized software platforms (Greenplum, HaDoop, R, SAS) to provide an efficient, scalable research computing platform for visualization, large database matching and analysis.
Each of the systems listed above connects to a centralized storage system providing over 130 TB of storage for researchers. Storage platforms include Isilon, EMC and Compellent.
PHS utilizes many specialized applications to perform statistical genetics research. The following list comprises a subset of these applications:
- Gene Hunter
- HRT - Haplotype Runs Test
- Non-Parametric Linkage Regression Analysis
- OSA - Ordered Subset Analysis
- PDT - Pedigree Disequilibrium Test
- Unphased (includes haplotypic PDT and QPDT)
- Dandelion - developed at Division of PHS and performs case/control haplotype analysis using the EM algorithm
- Dprime - developed at Division of PHS and calculates estimates of D, dprime, delta and r-square using independent individuals
- SNPGWA - developed at Division of PHS for genome-wide association analysis, performs linkage disequilibrium measures, tests for Hardy-Weinberg Equilibrium, tests for general association, 3 genetic models (dominant, additive and recessive), tests for lack of fit, performs an allelic test and performs 2 and 3 marker haplotype analysis using EM algorithm.
All resources are physically located in one state-of-the-art 5,500 square foot data center designed to ensure security and reliability. It is located at the new downtown campus of WFSM and provides a central presence for high-end computing (e.g., computing clusters), data storage facilities, centralized backup and disaster recovery.