Uncovering new ways to treat diseases from increasing biological data remains a daunting task for a life sciences researcher. With new drugs taking years of resource-intensive research, development, testing and marketing, researchers at pharmaceutical and biotech companies must ensure that processes are completed as quickly, accurately, and inexpensively as possible. Thankfully, high performance computing (HPC) is coming to their rescue by allowing them to analyse massive amounts of data for genome analysis and drug design more quickly and cost effectively.
The Council of Scientific and Industrial Research (CSIR) centre for high performance computing in New Delhi hosts a four-teraflop Hewlett-Packard supercomputer at its facility in New Delhi, running the Linux operating system to advance its life sciences computational biology research.
?This is the among the most powerful HP supercomputer in Asia and vaults the centre into the ranks of global research institutions such as Pacific Northwest National Laboratory, Pittsburgh Supercomputing Centre, Sandia National Laboratory and Los Alamos National Laboratory that have implemented scaleable, multi-teraflop supercomputing systems,? says HP India Enterprise Business marketing and growth initiatives head, Faisal Paul.
The scientific exploration ranges from genes to proteins and from biotechnology to pharmaceuticals and personalised medicine. ?They now possess multiple whole genomes to look for differences and similarities between all the genes of multiple species. From such studies, we can draw particular conclusions about species and general ones about evolution?often referred to as comparative genomics,? informs Paul.
Not far away, the supercomputing facility for bioinformatics and computational biology at the Indian Institute of Technology Delhi is serving 4,000 researchers worldwide to reduce computational time on complex research projects for genome analysis, protein structure prediction and drug design. The facility uses its high performance computing infrastructure for biomolecular modeling and creating new software used as tools for genome/protein structure based medicine development. The facility has an aggregate compute power of more than 600 Gflops and data storage capacity of 4 terabytes.
?We have taken on major challenges in genome analysis, protein folding and drug design research,? says the facility?s coordinator, B Jayaram. Going forward, the facility wants more insight into bigger problems such as annotating and categorising the human genome, automated protein structure prediction and lead molecule discovery.
Among those who have gained include LeadInvent, a spin-off company from the supercomputing facility at IIT Delhi. ?We combine proprietary computational biology technologies: computer aided drug discovery algorithms?SanjeeviniPro and protein structure prediction algorithms?BhageerathPro with high performance computing to develop viable therapeutic lead molecules,? says the company?s co-founder, Surajit Bose.
Down south, Centre for DNA Fingerprinting and Diagnostics (CDFD), Hyderabad has deployed Sun Microsystems? high performance computing systems to aid research in DNA fingerprinting, diagnostics and bioinformatics. University of Hyderabad has set up a high-end supercomputer infrastructure to support research activities in life sciences.
In Bangalore, Cellworks Research India is using high performance computing to develop new therapeutics with companion bio-marker based diagnostics for complex diseases like oncology, auto-immune disorders, type-2 diabetes and central nervous system disorders, specifically Parkinson?s. ?High performance computing allows much faster debugging capabilities and allows us to batch our combinational drug studies, like never done before. We are using high performance computing to test out the virtual protemics platform that allows us to do thousands of virtual biology assay/experiments in a matter of days, compared to traditional techniques that could take months to converge,? says Anand Anandkumar, chairman and managing director, Cellworks Research India.
Open source drug discovery (OSDD) initiative of CSIR is equally gung ho on high performance computing. Zakir Thomas, OSDD project director and head of CSIR director-general?s technical cell says, ?We are connecting three major high performance computing centres namely, IGIB, IMTECH and JNU through the National Knowledge Network (NKN). OSDD has established HPC infrastructure at these three locations which are connected by the high speed NKN. Providing a robust IT backbone for a community of 1,800 researchers would not have been possible but for the interconnected HPC infrastructure available at the three locations.?
It is evident that a high performance computing wave is sweeping across the scientific, research and academic landscape in India. These institutions are now building the high performance computing clusters for meeting their toughest challenges. Life-sciences companies and those institutions involved at the cutting-edge of life-science research face fierce competition and rising R&D costs as they work amidst various challenges to bring a breakthrough drug to market. Technology, especially in informatics and computational sciences has an important role to play in helping further research in life-sciences.
The IT infrastructure must align with constantly changing business and research requirements. High performance computing has also helped many pharmaceutical companies expedite drug discovery. Large simulation experiments such as molecular interactions/dynamics, protein folding, virtual drug screening, discovering RNA therapeutics etc. are completed with faster turnaround times with high performance computing. This technology not only accelerates discovery of new drugs and targets with minimum side effects, but also enables solving fundamental problems of gene regulatory network and host pathogen specificity.
This uptake of high performance computing by the life sciences industry will see an upswing with the government?s focus on research on affordable medicines. ?We believe that 2010 is the point of inflection when life sciences firms in India will start reaping the benefits of high performance computing. This will happen because of the adoption of high throughput computation techniques that has started to take wing this year,? says Anandkumar. The use of systems biology techniques and chem-informatics products to reduce the turnaround time associated with drug discovery, have hastened the need to put to use high performance computing, he adds.
High performance computing, which used to be synonymous with supercomputers, refers to the use of computers for solving complex computing problems, be it nanotechnology, 3D modeling, electronic design automation, intensive research and computer aided animation. This domain-intensive space uses high-end software and hardware to simulate real-life scenarios.
?Biological data has grown exponentially and query and data mining of this data cannot be done using normal computer workstations and require high performance computing,” says Thomas of CSIR. In his opinion, high performance computing requires a high capital investment and has a high depreciation rate. ?Normally, Indian pharmaceutical industry would like to be an user of this facility rather than being an owner of one,? he adds.
The high performance computing segment can be classified into supercomputers and clusters. The supercomputer that involves symmetric multi-processing systems (SMP)/ non-uniform memory access architecture (NUMA) is a computer that uses a large number of high-performance microprocessors. It is quite expensive.
Clusters on the other hand, provide the compute power by dividing the task among a number of computers. Terabytes of data can be processed over networks in seconds. This level of computing is what keeps business efficiency and productivity in enterprises running smoothly. High performance computing provides an edge to help solve highly complex problems, perform business-critical analysis or run computationally intensive workloads. High performance computing solutions vastly improve engineering computational work-flows by allowing disparate systems to be pooled and managed as a common computing resource. ?In the life sciences sector, high performance computing can accommodate this industry?s wide range of computational needs to solve critical challenges,? stresses Paul.
Typically, high performance computing usage helps in the in-silico analysis in drug research. If a larger part of drug discovery is done in-silico, including achieving the holy grail of systems biology, it will cut down the wet analysis and clinical research dramatically, thus reducing the cost of new drug discovery to a fraction of what it costs now. This is critical for drug research in diseases that kill many but do not have a significantly large market in monetary terms.
Jaijit Bhattacharya, director, government affairs, Hewlett-Packard India says, ?Drug companies are not able to invest in discovering drugs for such diseases as the economics does not support the research. This is where it is absolutely imperative to invest more in HPC to cut down on the total cost of drug discovery and also speed up the process of drug discovery.?
Globally too, pharmaceutical and biotech companies are hard-pressed to bring newer drugs to market faster. With new-found diseases coming to light everyday, researchers are trying to find answers by running extremely complex life-science simulations. High performance computing can accommodate this industry?s wide range of computational needs that are required to solve its problems.
Not surprising, researchers at one of the largest medical centres in the United States and a clinical research hub, Cedars-Sinai Medical Centre, are using a Sun Microsystems high-performance computing grid to make discoveries that could lead to personalised medical treatments for life-threatening and chronic diseases. The solution is being used to process and analyse vast amounts of complex data for a major medical research project to develop personalised predictive medical treatments to better manage chronic diseases and to help patients overcome life-threatening diseases. Early results have been encouraging. Two terabytes of data is analysed daily and complex data is analysed in days rather than weeks and months.
Similarly, researchers at the Laboratory of Neuro Imaging (LONI), University of California at Los Angeles (UCLA), are using high performance computing for treating and even curing a wide range of brain diseases such as Alzheimer?s and schizophrenia. LONI?s unique contribution to neurological research is the creation of digital atlases drawn from thousands of scans of individual human brains.
The associated collection of data is enormous, over 30 terabytes?and so are the computational challenges of analysing that data. But researchers are benefiting from a new high performance computing system and have significantly expedited their efforts for investigating the effect on the brain caused by schizophrenia and AIDS. This research could speed the development of new treatments, relieving suffering and saving lives.
At the Queensland Brain Institute in Australia, scientists are generating hundreds of gigabytes of research data as they investigate the more-complex areas of neurology. The research is vital to the development of therapeutic treatments for a range of degenerative neurological conditions such as dementia and motor neuron disease, as well as schizophrenia and depression. And over a year, this all adds up to terabytes of information that needs storing?a task recently met by an enterprise high performance storage system. The institute has gained unlimited data-storage capacity, reduced cost per terabyte by 65% and eliminated silent data corruption.
Closer home, as Indian pharmaceutical and biotech industries advance their technologies and expand their offerings, their computational requirements will increase tremendously. Additionally, in order to keep pace in a competitive market, they will need faster turnaround times for the R&D and production computational projects. Thankfully, high performance computing is helping them meet these demands.