Harvard researchers fold proteins with D-Wave quantum computer
December 27, 2012
Many were skeptical when, back in 2007, Canadian company D-Wave announced that it had built the world's first commercially viable quantum computer. Now a study published in the August issue of Nature's Scientific Reports co-authored by D-Wave and Harvard researchers proves the D-Wave One is the real deal.
"The D-Wave computer found the ground-state conformation of six-amino acid lattice protein models. This is the first time a quantum device has been used to tackle optimization problems related to the natural sciences," said Harvard professor Alán Aspuru-Guzik, the lead author of the paper.
The precise details of the protein-folding study are a bit complex, but basically they were looking for the lowest-energy configuration of folded proteins, which is believed to be the correct one since it is the most stable. In nature, proteins fold themselves correctly most of the time, but when they don't they cause diseases such as Alzheimer's. The quantum computer correctly solved 13 times out of 10,000 for four-amino-acid and six-amino-acid sequences under the Miyazawa-Jernigan model of lattice protein folding.
In the abstract, the authors seem optimistic about the computer's prospects, stating that, "the approach employed here can be extended to treat other problems in biophysics and statistical mechanics such as molecular recognition, protein design, and sequence alignment." And Google has adopted the system to train image recognition software.
What's next?D-Wave has been working on a 512 qubit chip since 2011, but since publication lags behind current work, this study used the company's 128 qubit chip. The experiments only required between five and 81 qubits, and while the 128 qubit chip isn't as powerful as regular supercomputers (or even some desktop computers), the speed is determined mainly by the type of quantum algorithm being run.
However, one can extrapolate speeds, and in an interview with NextBigFuture, D-Wave CTO Geordie Rose said that 512 qubits would be approximately 1,000 times faster than 128 qubits, and a projected 2,048 qubit chip would be 1,000 times faster than 512 qubits.