High performance computing (HPC) is often recognised as having made major contributions to industries like aerospace, advanced manufacturing and finance, but for many years companies in the life sciences industry have also been making aggressive investments in HPC technology.
To date, HPC has contributed to the advancement of life sciences in a number of important ways, helping to extract value from the massive data volumes generated by the “omics” fields of genomics, proteomics, and metabolomics, as well as advancing modeling techniques that have improved our understanding of biomolecular reactions, diseases, and organ function.
One of the most visible examples of how HPC has benefitted life sciences is in the sequencing of the human genome. HPC has long been a driving force behind human genomic research, and this will be increasingly true as the thirst for human genomic information continues to grow. The push for better genomic data has already reduced the cost of sequencing an individual human genome from millions of dollars to $1,000. Thanks to the increased availability of HPC, the cost could soon be reduced to the hundreds or even tens of dollars range, with the hope that individual genome mapping eventually becomes standard practice. The ready availability of genetic testing has and will continue to have a revolutionary effect on a variety of disciplines within the life sciences, providing better health surveillance plans for adults with a history of illness in their family, screening tests for newborns to prevent the development of dangerous genetic diseases, and new genetically tailored therapies.
In fields like oncology, HPC systems are being used to not just gather this data, but process it in novel ways that are leading to more effective treatments. Research into signal pathways and their relation to tumor growth is one of these areas. Signal pathways are a group of molecules in a cell that work together to control cell function, such as cell division and cell death, by passing signals from molecule to molecule in the cell until the cell function is carried out. Researchers hope that by better understanding these chains they may locate the signals that cause cancer, and learn how to inhibit them. The process of signal pathway research includes collecting tumor samples and genomic information, then correlating the interaction between tumor cells and treatment drugs to determine the most effective therapies. This computationally intensive process relies on applying data mining and machine learning algorithms, two typical HPC applications, to large genomic information datasets.
Other scientists have taken a very different approach to oncological research using HPC. Instead of studying tumor tissue, which may obscure the actual cause of the cancer, other research groups have taken a single-cell approach to studying the development of tumors. By using mathematical models to simulate cells, manipulating variables such as hormone levels, and observing the changes in the cell that take place, they aim to determine the precise probability of tumor development. This method, which requires the dissection of hundreds of cell mutations, would also be impossible without the advanced modeling capabilities that HPC enables.
Is it also worth noting that Artificial Intelligence (running on HPC) is being used for mammogram studies, significantly shortening the diagnosis phase and allowing treatment to commence sooner rather than previously possible.
Oncology is just one of the ways in which HPC is now being utilised within the life sciences. Other areas of great progress include the field of epidemiology where, for example, complex models are being created to plan the control of pandemics, and potential threats from biological weapons. In the neurosciences, where there is significant energy being invested in super-resolution mapping of the neuronal circuits of the brain, HPC is helping to not just harness the vast amount of neurophysiological data being produced by experimentation but to process intensive simulations with the resulting data.
Despite this progress, there are obstacles to further HPC adoption by the life sciences industry. Many life sciences applications are developed for desktop computers or small compute clusters, meaning that HPC capabilities are still not being fully applied in the field. Compounding these technical challenges is the fact many scientists and researchers working in the life sciences industry lack the IT training required to be comfortable with the operation of HPC systems, such as the command line, which further complicates adoption.
Adding to these problems is the fact that many research institutions lack the hardware resources to fully exploit the power of HPC. One human genome contains between 300 to 500 gigabytes of unstructured scientific data. This means that processing the data for a group of people—let alone an entire population—would quickly reaches proportions beyond the ability of normal IT systems. The enormous growth of biomedical data as a result of these new sequencing technologies means that institutions must be well-resourced with massive storage and computational capability that can only be found in HPC systems.
Verne Global considers itself a long-term partner to the life sciences industry. By providing flexible and convenient carbon-neutral HPC compute and storage capability to our partners in the life sciences, we not only put the power of HPC within reach of more researchers and scientists, but also significantly lower their HPC power cost so they can spend more money on research. We’re very proud of the 70% energy savings that the Earlham Institute has gained by partnering with Verne Global, for example. We hope the many benefits of partnering with Verne Global will expedite the increased adoption of HPC in the life sciences and assist advancements in vital areas of scientific research.