As quantum computing passes a “precarious milestone”, what are the implications for life science research?

It’s been hard to miss the coverage in the scientific and popular press over the last few days about Google’s claim (published in Nature ) to have achieved “quantum supremacy”. Without getting too deep in the weeds, Google’s paper is interesting to computer science experts, but the reality is we are still a long way off from a working quantum computer. When a working model does arrive, however, what will it mean for those of us in the life sciences?

A traditional computer uses bits in a binary system – 0’s and 1’s – to model any kind of data imaginable. Quantum computers use “qubits,” which can be made of a variety of materials and which exist in a sort of limbo between zero and one until they are observed. So while a traditional bit can be only a zero or a one, a qubit can be a zero, a one, both, or anything in between. Each qubit added to the computer exponentially increases the data processing power of the system.

In practical terms, a traditional computer can only process one computation at a time. For example, if you are trying to guess a password on a traditional computer, you could write a program to try one password at a time until you get the right one. A quantum computer does not have this restriction, and with enough qubits, you could hypothetically try every password in the universe simultaneously, arriving at the answer in seconds. The speed of each individual computation is the same for quantum and traditional computers – the difference lies in the number of computations that can occur at once.

For everyday computer uses this computing style is less efficient than traditional models, and in fact would be slower for the most popular computer tasks (watching videos, writing documents, browsing the Internet, etc.). Quantum computers are also very expensive to build and run, and take up entire rooms like the traditional computers of old. However, considering the example above, it is easy to understand why most governments and many corporations are investing heavily in quantum computing research.

So far, so “Star Trek”. But what might the impact of quantum computing be on the life sciences and, in particular, drug discovery, which is our passion here at Charles River? There are several companies who are pursuing quantum computing approaches in this field and surveying their reported activities may give us some clues.

Perhaps unsurprisingly, some are looking to see how quantum mechanical (QM) calculations of molecular systems (which are already possible, but demanding, on conventional computers) might be advanced through quantum computing. Just a few months ago, Schrödinger – a molecular modelling software developer – announced a collaboration with Qu & Co, a quantum-software development firm. Few details are available, but the initial focus of the collaboration appears to be in the realm of materials science, but the QM calculations and “atomistic simulations” mentioned are equally applicable to drug discovery.

ProteinQure is a biotech firm based in Toronto that aims to perform drug design in silico. As part of that effort, the company is exploring quantum computing. In a white paper, entitled “The Quantum Landscape”, the company suggests that several key molecular modelling tasks may be accelerated by hybrid classical/quantum computing approaches in the not-too-distant future. In particular, there is interest in “quantum annealing” (as implemented on D-Wave systems) which can offer “tunnelling” through the energy barriers that hinder traditional search and optimization algorithms and thus speed the identification of optimal solutions to complex problems such as protein folding and design, and protein-ligand docking.

Another core molecular modelling task is the comparison of one potential drug molecule with another, i.e., the calculation of what is known as “molecular similarity”. More than two years ago a proof-of-principle study was announced by another specialist in quantum computing, 1QBit, working together with Accenture Labs and the biotech company Biogen. The approach has been published (as a pre-print) and involves a novel graph-based approach for determining molecular similarity which is amenable to solution using a quantum annealer. According to the study, the quantum computing method allowed Biogen to compare molecular structures more accurately and to obtain more information about the comparisons than any of its existing, classical computing-based methods.

More recently, Qulab has posted results (again as a pre-print) of new research focused on molecular dynamics simulations and quantum computing technology. One of the issues limiting the reliability of molecular simulations is the upper bound on accuracy imposed by the current generation of molecular mechanics-based force fields. The Qulab research looks at how the development of new force fields using scalable ab initio quantum chemistry calculations on quantum computers can help to address this limitation. Based on this work, the company believes that “near-term” quantum computers (comprising ~300 qubits) will have tangible benefits on the drug discovery process in the next three to five years.

Of course, this survey is not exhaustive and it is worth remembering that these companies all have a vested interest in quantum computing. Nonetheless, it does give some idea of the potential application areas for this novel computing technology and a timeframe for it to begin having an impact on drug discovery.