Advances in quantum computers could transform the way we discover and understand new drugs.
There are established methods of predicting and understanding the effects of potential drugs at a molecular level. Molecular dynamic simulations use computational simulation of the interaction between atoms in a molecule over time.
Using this approach, we can understand how a molecule would bind to a protein and affect its structure, and through those changes its function.
This method is often augmented by density functional theory and involves calculation of the energy of a molecule to predict structure and how that molecule will interact with other molecules.
These approaches require huge and costly computational power, and running these simulations takes a long time even with today’s most powerful computers.
Compared to classical computation, quantum computation provides a massive and revolutionary increase in computational capacity with a way of processing information that is beyond what classical physics allows, opening new opportunities to understand micro- and macro-level events.
Classical computation is built upon ‘bits’, which are either 1 or 0 depending on whether current is flowing through a transistor. A quantum bit, or ‘qubit’, is a superposition of two states at the same time.
This means that it can be both 0 and 1, which is not the same as it being 0.5!
In a quantum state the bits are running multiple computations simultaneously, until an ‘answer’ is found, at which point they ‘decohere’ and you get classical bits as your output. That makes them incredibly fast problem solvers.
Classic track?
Classical computers cannot perfectly model quantum systems because they obey the laws of classical physics.
Molecules and their interactions can only be approximated by classical computational methods irrespective of how much computational capacity you have.
Classical computation and quantum computation are differentiated by the fact they do things differently, not simply at different rates.
By way of example, in the pharma industry the use of hierarchical modelling is a sequential process like that of Darwinian evolution, where at each decision point we ask, ‘is molecule A better than molecule B?’
You might be led down the garden path, however, as ultimately the decedents of the less well-matched molecule might prove better than those of the alternative further down the road.
Quantum computers would work on solutions in parallel such that there are no early-stage decision points. This is particularly powerful when tens of thousands of compounds need to be screened.
Next frontier
Paradoxically, we might find out in a not-too distant future that the AI platforms we have developed for drug discovery are not equipped to take full advantage of quantum computers.
That today’s AI drug discovery platforms become obsolete with the emergence of quantum computers.
Quantum mechanics has a long and complex history of disputes as to its underpinnings and our capacity to harness its potential power. What is clear, however, is that interest and funding continue to grow in the next frontier of technology.
Dr Joe Taylor is Principal at Candesic and Dr Leonid Shapiro is Managing Partner at Candesic. Go to candesic.com
