eRSA services in use: HPC, Storage & Cloud, Big Memory nodes
Dr Andreas Schreiber’s work with the ACRF Cancer Genomics Facility has made significant advancements in combatting the deadly type of cancer known as Chronic Myeloid Leukaemia (CML). Operating out of SA Pathology’s Centre for Cancer Biology and in collaboration with the University of Adelaide, Andreas and his team have worked alongside eRSA since 2012.
Originating from a genetic mutation referred to as chromosomal translocation, the cause of CML is well known among researchers. CML starts when this mutation occurs during cell divisions in the body, which results in one chromosome connecting to another. When the chromosomal breaks happen to lie in genes, as is the case with CML, a fusion-protein is created that should not exist, causing cancer.
Promisingly, research into cancer genomics has resulted in drugs designed to attack this aberrant protein, stopping its action without having to use chemotherapy. Prior to the development of these drugs, receiving a CML diagnosis was a death sentence. Since patients have started taking these drugs in the last 10-15 years, people have continued surviving, creating the enviable situation of researchers struggling to determine the altered life expectancy due to its effectiveness in treating the cancer.
However, for a certain fraction of people, the drugs are not effective in controlling the progression of the disease. In this instance, Andreas’ team is able to use their gene sequencers to investigate the cause of the failure. Ultimately, the goal is to use this information to tailor treatment to the individual patient. Andreas refers to this process as an example of “personalised medicine”, where the medicine is designed and prescribed specifically for the patient.
Sequencing can be a very time-consuming task, but Andreas says best practices have improved to be more efficient. Instead of analysing mutations in solitary genes one at a time, researchers are now performing Whole Exome Sequencing. This technique examines all 20,000+ protein-encoding genes within a genome (the complete set of genetic material within a cell) at the same time. Once presented with this data, medical experts are then able to identify patterns and mutations in particular genes that may be leading to disease. Clinicians can use this information to determine what specific follow-up testing to employ, as Andreas explains.
“Traditionally, you might have had a child that has a disorder or cancer, they get tested for mutations in a gene associated with the illness, the clinician will get that tested and often it comes back negative,” Andreas said. “The clinician will then send it back for another test for mutations in another gene.”
“And this can go on for a long time, which is referred to as a ‘Genetic Odyssey’. With Whole Exome Sequencing, you can get everything tested at once and find out straight away.”
This does come at a cost. The output of genome and exome sequencing creates terabytes upon terabytes of data that simply cannot be stored on a desktop computer or on an on-site server. Andreas works with eRSA to ensure all sequencing data is stored and accessible remotely, and is securely backed up as a safeguard. With continual advancements in sequencing technology, an increasingly higher volume of data is being produced within ever-decreasing periods of time. In addition to storing this data, a high amount of processing power is required to analyse the large datasets, which eRSA’s HPC Big Memory ‘Big Mem’ Nodes are equipped to handle.
“We are analysing more and more genomes which is something we hadn’t done before,” Andreas said.
“For each genome, we now have 50 times more data to process than for an exome, and we couldn’t do this with our old server. Big Mem is a scaled-up version of our own server. With Big Mem, we use lots of cores and using between 20-30 gigabytes of memory per core. We want to do everything in the same amount of time, therefore requiring parallel processing capability.”
eRSA eases the processing and storage load on Andreas’ team via eRSA’s new HPC platform, TANGO. This hardware is needed because of the greater sequencing complexity which Andreas states is due to the rapidly increasing technology advancements and decreasing costs associated with sequencing.
“The cost of sequencing often gets compared to Moore’s Law (where computer processing power doubles roughly every two years),” Andreas said.
“The cost of computing halves every two years, and the cost of sequencing is decreasing even faster than that, which means more and more people can afford to sequence.”
“Years ago, the first genome took an international collaboration of hundreds of scientists more than five years, $3billion, and factory-floors full of old-style automated sequencers. Now, one genome gets sequenced by one lab in one day for a little over $1000. A whole exome costs a few hundred dollars. This is in part thanks to the services eRSA provides”
“In addition to this, eRSA’s infrastructure helps us look back on data that we took years ago that’s still relevant today – having the data still there, it’s just perfect.”
Andreas enjoys working alongside eRSA, citing easy communication and great problem solving processes as some of eRSA’s strong points.
“They (eRSA) know us quite well,” Andreas said.
“The communication with eRSA is great – they not only have a strong understanding of our needs, but they are always readily available to problem solve any issues that arise and always keep us well informed of any updates or upgrades. We really just appreciate the way they work.”
Want to talk to Andreas about his research and the tools used?
Phone: 08 8222 3965