Researchers from the University of Oxford, Roche and UCB Pharma, have reviewed how the faster algorithms promised by quantum computing may affect everything from data and network analysis to electronic structure simulations and protein modelling.
In the article published last Friday in WIREs Computational Molecular Science, the authors examine how quantum computing will likely shape the future of computational biology and bioinformatics, by looking at the benefits and limitations of the technology will bring, with a nod to the inevitable hype it has had.
Despite great progress in modern computing capabilities, many challenges, particularly in the remit of protein folding prediction, calculations of binding affinity, or selecting optimal large-scale genomic alignments, remain infeasible to today’s supercomputers. Quantum computing, a more powerful generation of computing, may hold the solution to these challenges, and much more. The difference between the technologies is not in their processing power but rather a paradigm shift in their operations, whereby algorithms of quantum approaches can achieve unprecedented speeds in computational tasks.
The authors describe one such example of a quantum speed-up: “computing the full electronic wavefunction of an average drug molecule numerically is expected to take longer than the age of the universe on any current supercomputer using conventional algorithms, while even a modest‐sized quantum computer may be able to solve this in a timescale of days.”
However, we have not yet reached the dawn of quantum computing, for the various technical roadblocks are staggering. These include difficulties related to the manufacturing of hardware, to controlling and limiting the noise for quantum systems.
After introducing quantum computing as a concept, as well as quantum information processing, the authors focus on three areas of computational biology where quantum computing has shown the greatest algorithmic progress: statistical methods and machine learning, electronic structure calculations (quantum simulation), and optimisation problems (including protein structure predictions).
The authors finally discuss the medium- and long-term impacts on computational biology, that quantum computing may bring: “The potential of even small quantum computers to outperform the best supercomputers on certain tasks may prove transformative to computational biology, promising to make impossible problems difficult, and difficult problems routine.”
Journal reference: The prospects of quantum computing in computational molecular biology
C Outeiral, M Strahm, J Shi, G.M. Morris, S.C. Benjamin, C.M. Deane