About KCCG



A high performance computing environment

The KCCG high performance computing environment has been tailored for genome informatics. The system, housed within a Garvan IT data centre, comprises almost 2000 CPU cores, with 512 GB of memory per node and 100 Gb network switches. Our newer generation (April 2016) nodes each have 20 TB of non-volatile memory (NVMe) that enables extreme performance on the node, with our parallel file system providing fast transfer to and from the node.

To embrace changing models of computation we support bare metal HPC (Rocks), OpenStack Cloud and bare metal Hadoop and Spark clusters. KCCG runs extensive quality control (QC) tests on each sample we sequence to ensure that we meet all our commitments around sequence quality and coverage. These QC analyses are run on our local infrastructure.

The rest of our infrastructure is for rapid prototyping and interactive analyses, but not for scaling into production. Instead, once locked down analyses are run at scale at one of our partnering organisations that include the National Computational Infrastructure (NCI: nci.org.au), an Australian supercomputing facility as well as AWS and DNAnexus. We have a direct connection to AWS Sydney (provided by UNSW Sydney) and a dedicated high-speed link to the NCI that we also use for data archiving.


A collaborative research centre

KCCG undertakes collaborative research projects with scientists at the Garvan Institute as well as external partners, in all areas of medical research. The Centre also undertakes research to improve the interpretation of genomes and genome variants with the overall aim to further the use of genomic information in patient care. 


Aiming to close the gap between research and the clinic

Close partnerships between research institutes and hospitals further genomic research though better translation and transfer of information from research to clinic, and from clinic to research. Our clinical and research collaborations are key to realising the benefits of genomic medicine in Australia.


Developing infrastructure for storage and analysis of genomic information

The computational challenges in integrating genomic and clinical data require significant investment in storage and compute capacity and software infrastructure. Work at KCCG involves designing a roadmap for the future of genome data analysis for the Garvan Institute and our collaborators, making use of offsite resources, while maintaining patient sample security and integrity. Through the combination of clinical-grade analysis and informatics research, the Centre is creating an environment where a critical mass of bioinformaticians – including researchers funded by external groups and a wide variety of research groups within Garvan – are co-located.


Engaging communities and stakeholders in genomic medicine

If clinical genomics is to provide benefit for the Australian public as a whole, the patients, professionals and the diverse communities that make up our population will need to be engaged. Well-designed and integrated public and professional education efforts are required to address this complex and fast-moving field.

A purpose-built facility for undertaking genomic sequencing and analysis

The Kinghorn Centre for Clinical Genomics was one of the first sites in the world to acquire the Illumina HiSeq X Ten sequencing platform, which the powerful genome sequencing capability of sequencing up to 18,000 whole human genomes per year. Research genomics and data analysis is now available through Garvan's wholly owned subsidiary, Genome.One.


Clinical accreditation for its sequencing, analysis and reporting

The drive behind KCCG is the of return of accurately interpreted clinical-grade sequencing information to physicians to inform patient care. KCCG, through Genome.One has achieved ISO15189 (medical testing) accreditation for its processes from blood sample to  clinical report. Our accreditation is currently limited to testing for genetic diseases, which includes rare and inherited conditions. Clinical whole genome sequences are available through Garvan's wholly owned subsidiary, Genome.One.


Header image credits: P. Morris/Garvan