A new treatment for Zika virus could soon be brewing inside average computers and mobile phones. Virtual experiments, run on a globe-spanning supercomputer, are already blazing through millions of compounds against key target proteins. Once the field is winnowed, robots will test likely contenders.
Brazilian professor Carolina Horta Andrade is leading the ambitious project, called Openzika. Using IBM Corp.'s free World Community Grid, her team is harnessing spare computing cycles to run virtual screening tools. The work is part of a global effort that already includes at least 67 companies and research institutions working on diagnostics, vaccines, therapeutics and vector control. With the virus already found is 60 countries including the U.S. and expected to spread to Europe this summer, it is desperately needed.
"As a scientist and a citizen of Brazil, which has been greatly affected by Zika, I am committed to the fight against the virus," Andrade recently wrote. Using only the resources of her lab at the Universidade Federal de Goiás, the team would only be able to screen a few thousand compounds against a subset of Zika proteins in a reasonable period of time. With the World Community Grid, they will evaluate more than 20 million compounds.
The grid is comprised of several hundred thousand active devices on average, Viktors Berstis, IBM's chief technologist for the Grid project, told BioWorld Today. Altogether, it puts the power of 500 to 600 CPU years at researchers' disposal each day, the equivalent of 500 to 600 computers, each running all day for one year. Since its inception, the Grid has contributed the equivalent of running a single decent computer for 1.2 million years to a variety of projects, including one to identify new materials for more efficient solar power cells and another led by the Chiba Cancer Center Research Institute that led to the discovery of seven drug candidates that could potentially be used in new medicines to fight childhood neuroblastoma.
Scientists tend to "instinctively reduce the scope of their projects to what's doable with the resources on hand," said Berstis. They never entertain the idea of what they could do if they had a fantastic amount of supercomputer time, he said. Though some scientists already work at universities with national supercomputing centers, they compete with colleagues for the resources of those centers. On the World Community Grid, they gain a more powerful and capable share of the available computing cycles. "We like all these projects because every time [new investigators] come on board, they suddenly have this a-ha moment where they realize they've just gone several orders of magnitude higher in what they thought they could achieve," he said.
The Grid is built using BOINC, the Berkeley Open Infrastructure for Network Computing, an open-source grid computing system originally developed at the University of California, Berkeley, to support the SETI@home project, a program that enlisted volunteers to run software capable of downloading and analyzing radio telescope data as part of a search for signs of extraterrestrial intelligence. To perform its computational experiments, Openzika researchers are employing a widely used virtual screening tool called Autodock VINA, developed by the Olson laboratory at the Scripps Research Institute.
The Openzika project grew out of a 2015 project that Andrade worked on with pharmacologist Sean Ekins, CEO of Collaborations Pharmaceuticals Inc. Together they had been working to develop computational models to identify active compounds against the dengue virus. In January 2016, the colleagues decided to expand their dengue research to include Zika, since the two diseases are from the same family of viruses, flavivirus. A month later, they published an article co-authored with others in F1000Research calling for an open drug discovery effort. That idea is now embodied in large part by the Openzika project.
One early contributor to the project who is already running the World Community Grid software is Jair Siqueira-Neto, who was born in Brazil. But that's only the start for Siqueira-Neto, an assistant professor at the University of California, San Diego's Skaggs School of Pharmacy and Pharmaceutical Sciences. Once the computer modeling phase of the project identifies promising candidates, he'll enlist the help of robotic liquid handlers and a robotic microscope at the university to rapidly test between tens and hundreds of thousands of candidates identified by the project against real-world Zika virus. "It's really an approach that matches the two technologies well," he told BioWorld Today.
He expects the first batch of candidates to be released within the next few weeks or months. "With Zika in my home country, Brazil, I really wanted to contribute," he said. "We can really accelerate the whole drug discovery process, taking the virtual combined with the actual real experimental approach. That's really key for this project," he said.
Although deeply dependent on the generosity of individuals ready to donate space computing power, the project also rests heavily on infrastructure provided by the Grid, a philanthropic initiative of IBM. The real-life lab work at UCSD will also require non-free resources, such as reagents and other supplies for testing and, eventually, chemists to make modifications to optimize activity of the most promising molecules. Siqueira-Neto said that such work will most likely be funded by requests for grants from the National Institutes of Health.
"As science's knowledge of the Zika virus increases in the coming months and key proteins are identified, the Openzika team will use the new knowledge to refine our search," wrote Andrade. Supporters might also take note of another of the award-winning scientist's smaller goals: "I want a #Openzika T-shirt," she tweeted recently.