India’s Department of Biotechnology and Sun Microsystems have set up the country’s first supercomputing facility exclusively for bioinformatics and computational biology, calling it a first step towards a national biocomputing grid.
Biologists say the facility will provide Indian researchers computational resources to address real-world challenges, but some cautioned that talk of a grid is premature for the moment.
The new facility at the Indian Institute of Technology, New Delhi, hosts a 70-processor Sun cluster, a 16-processor Pentium cluster, and a 4-processor Silicon Graphics Origin 200 machine with an aggregate computing power of 150 gigaflops.
“We expect this infrastructure to enhance India’s contributions to genome analysis, protein folding, and in silico drug design,” said Manju Sharma, India’s biotechnology secretary.
The Department of Biotech-nology has indicated that it will spend $20 million on bioinformatics infrastructure and research over the next five years. The IIT facility received $200,000 from the department and a matching grant from Sun Microsystems. The department has also signaled its intention to set up another supercomputing facility for biology at the Indian Institute of Science.
Biotechnology funds have supported dozens of bioinformatics centers across publicly funded laboratories and universities across India since the early 1990s. But activities there have been restricted to establishing mirror sites of public domain databases, database analysis, and efforts at software development.
“A lack of computing resources has prevented researchers in India from addressing an entire class of problems that involve molecular dynamics, simulations of large biological molecules, and drug-target interactions,” said Manju Bansal, director of the Institute of Bioinformatics and Applied Biotechnology, Bangalore, and a faculty member at the Indian Institute of Science, Bangalore.
“There was an urgent need to complement ongoing bioinformatics efforts with high-end computing resources to do more meaningful research,” he added.
“The infrastructure we had until now could take us to a point, and no further,” agreed Bhyravabhotla Jayaram, principal investigator at the IIT. “Now we can look forward to working with atomic-level-detail simulations.”
Computational biochemist Jayaram and his colleague Achintya Das have developed new gene-to-drug software modules for protein sequence analysis, protein structure generation, selection, and optimization, and drug-target interactions. They describe their effort as a combination of bioinformatics tools and molecular dynamics to find short cuts to protein structure determination.
“The IIT, New Delhi, methodology and software seems to be able to determine certain kinds of protein structures faster than standard tools,” said Indira Ghosh, a senior research scientist with AstraZeneca’s Research and Development Center, Bangalore. Ghosh said his team plans to use the IIT modules and facilities to evaluate hypothetical proteins on Mycobacterium tuberculosis in its work to find novel targets to cut the duration of tuberculosis treatment from the current six to nine months to less than four months.
Biotechnology officials have said the IIT facility is expected to become a node of a national biocomputing grid. But researchers caution that while a biocomputing grid is a goal worth pursuing, there is currently far too little research in India on the application side of the equation.
“In addition to investing in infrastructure, significant efforts will need to be put into developing grid-enabled software,” said Dheeraj Bharadwaj, a computer scientist at IIT, New Delhi.
“Biologists will also have to get familiar with grid computing,” said Rajendra Joshi, a scientist at India’s Center for Development of Advanced Computing, Pune, which has developed India’s most powerful parallel supercomputers.
The CDAC has sought government approval for a $30 million supercomputing grid linking seven cities, listing “computational biology” among major application areas.
— Ganapati Mudur, New Delhi