Harvard researchers fold proteins with D-Wave quantum computer

Harvard researchers fold proteins with D-Wave quantum computer
The D-Wave One quantum computing system (Photo: D-Wave)
The D-Wave One quantum computing system (Photo: D-Wave)
View 2 Images
D-Wave processors on a wafer (Photo: D-Wave)
D-Wave processors on a wafer (Photo: D-Wave)
The D-Wave One quantum computing system (Photo: D-Wave)
The D-Wave One quantum computing system (Photo: D-Wave)

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.

Source: Nature via D-Wave

Raymond Johnson
Yay for Quantum Computing!
13 times it succeeded out of 10,000 tries? Hmm, makes Microsoft's windows look promising. I'll be impressed when it solves problems 9,999 times out of 10,000. i can live with a little tiny error.
Jess Atwell
Hate to be a downer, but this doesn't seem to be any improvement to current computing. And the thought that all that has to be done is increase it's computing power doesn't seem to be any different than adding power to current where will this take us in the end?
Chuck Anziulewicz
Maybe these new quantum computers will figure out a way to fold space. Then we can get on with the business of interstellar colonization.
I wonder how many times out of 10,000 tries a monkey would get it right? Just wondering if it statistically significant.
Yes Bew, what exactly is the success rate of a monkey solving six-amino-acid folding sequences under the Miyazawa-Jernigan lattice-model?
A working, commercially available quantum computer is a big deal. It is long way away from mass market but this is still a significant benchmark.
In normal computing there are two states, 0 or 1. In quantum computing there are multiple possible states or quantum bits (qbits) so it has the potential to scale exponentially.
It would have been helpful if the author had posted what a purely random approach would have yielded so we can gauge what the significance of 13/10000 is, since otherwise it looks lousy.
@Adrien the short answer is 0/10,000 but check out folding@home (, it is a distributed computing platform that uses mostly donated CPU cycles to complete protein folding.
Here is a quote from the page:
"It's amazing that not only do proteins self-assemble -- fold -- but they do so amazingly quickly: some as fast as a millionth of a second. While this time is very fast on a person's timescale, it's remarkably long for computers to simulate. In fact, it can take about a day to simulate about a hundred nanoseconds (1/1,000,000,000 of a second). Unfortunately, many proteins fold on the millisecond timescale (1,000,000 nanoseconds). Thus, it would take 10,000 CPU days to simulate folding -- i.e. it would take 30 CPU years! That's a long time to wait for one result"
If you are just adding a couple numbers in a single computation it is possible for a random result to occasionally equal the correct result but for something as incredibly complex with as many calculations required to complete a protein folding sequence the odds of a random result being correct are abysmal.
Ken Dawson
In other news: "What?!! Quantum computers are real!?!"
Ulf Lindroth
I think 13 out of 10,000 means the 13 correct minimum configurations out of 10,000 possible non minimum configurations. It doesn't mean 13 correct out of 10,000 correct.
Load More