Compute infrastructure and informatics know-how for Genomics
Computational Solutions Shared Resource is based in GCB's Informatics group; it operates, manages and maintains high-performance infrastructure tailor-designed for the compute, networking, and data moving challenges arising from genomic science. This includes a high-performance computing (HPC) cluster (HARDAC) with large parallel file system storage attached, OpenStack, and VMware-based compute VM clusters. Our infrastructure is used daily by researchers across schools and departments, as well as by cores within and outside of GCB. We maintain strong cybersecurity to meet the data protection requirements for PHI and other sensitive data.
Computational Solutions staff is continuously at the forefront of identifying and evaluating emerging trends and technologies in HPC for genomics. GCB's Informatics team includes scientific software architects and programmers who collaborate with GCB's faculty and core facilities on an ongoing basis, allowing us to regularly review, adapt, and optimize our services to ensure they meet the computational needs of next-generation genomic science.
We are adamant about empowering scientists to take control of their data, computation and scientific reproducibility. GCB Informatics staff also actively co-develop and co-teach data carpentry, reproducible science and scientific computing training bootcamps, and we are developing A Best Practices for Effective HPC training course. These courses are advertised and open to everyone at Duke and have been attended by a wide variety of learners.
We firmly believe that openness, transparency and reproducibility are cornerstones of science, and better, more reproducible computational infrastructure enables more and better science. Our computational infrastructure is Linux-based with strong commitment to open-source software. Genomics as a data-rich discipline gains tremendously from interdisciplinary collaboration, and we aspire to not only deliver a service but to be collaborators.
We aim to deliver high-performance computing and data management infrastructure designed specifically to meet the computational needs of genomic biology. For needs that can be met by general-purpose IT infrastructure, we interface with the highly capable commodity computing services available on Duke's campus.