CHICAGO – The Cleveland Clinic and IBM are taking both a short and long view in terms of quantum computing to uncover new insights in virology and immunology.
The two entities announced last week that they had reached a 10-year agreement to establish the Discovery Accelerator for healthcare and life sciences research using high-performance cloud computing, artificial intelligence, and quantum computing.
The Discovery Accelerator is essentially serving as the data center for the Cleveland Clinic's new Global Center for Pathogen Research & Human Health. Colloquially called the Pathogen Center, this establishment is focused on understanding viral pathogens, virus-induced cancers, genomics, immunology, and immunotherapies.
"Fundamentally, the Discovery Accelerator will be the technology foundation to … enable the mission of the Pathogen Center," said Anthony Annuniziata, director of the IBM Quantum Network. "It comes from a very simple desire to accelerate scientific discovery, to accelerate the progress toward new treatments [and] new solutions to these significant challenges."
The IBM Quantum Network is a consortium of IBM customers, including major corporations, startup businesses, universities, and national research laboratories, with an interest in quantum computing.
Annuniziata called the accelerator a bit of a new approach, in that it will rely on both established and emerging technologies, from artificial intelligence, high-performance computing, and hybrid cloud platforms to generative modeling — a somewhat new approach to machine learning in which computers automatically develop models based on pattern analysis — and the future centerpiece, quantum computing.
Annuniziata said that quantum technology right now has to work in concert with "classical" computing. "When you approach a whole-genome level of analysis, there is still going to be lots of stuff that's done by a classical computer and there'll be some things that are accelerated very significantly, we expect by a quantum computer," he said.
For example, an IBM spokesperson explained, modeling chemical reactions such as drug metabolism eventually can be split between traditional and quantum computers.
"It would make sense to model the actual reaction of the breakdown of the drug by the appropriate region of the protein with very sophisticated levels of quantum chemical methods," the spokesperson said via email, though this is likely not possible on current quantum computing platforms. "The rest of the protein would not be modeled on the quantum computer but rather with more approximate methods on a classical computer."
The Cleveland Clinic launched the Pathogen Center February with the help of a $500 million investment that includes support from the state of Ohio and JobsOhio, a nonprofit economic development corporation that promotes business in the state. The investment is intended to spur innovation and create 7,500 jobs in Northeast Ohio over the next 10 years.
"It's a team of researchers that study pathogens all the way from their basic biology to things like how do they infect humans? How do they interact with the environment? How does all of that translate into pandemics and public health emergencies?" explained Lara Jehi, the clinic's chief research information officer.
Jehi said that she expects the Pathogen Center to "put Cleveland and Ohio at the epicenter of research and innovation" for how pathogens and viruses influence human and public health.
As part of the Discovery Accelerator, IBM will install the first private-sector, on-premises instance of its IBM Quantum System One commercial quantum computing platform on the main Cleveland Clinic campus. IBM also said it would later build the first instance of its next-generation, 1,000-qubit quantum system there.
Annuniziata explained that quantum computing is a form of information processing that follows the laws of quantum mechanics rather than the ones and zeros from traditional binary computing.
He said that Quantum System One will include research first to understand the full capabilities of quantum in bioinformatics, then to build approaches that take advantage of those capabilities to answer specific research questions. The 1,000-qubit successor will "be an inflection point of sorts in the technology where we start to [be] able to handle larger-scale, practically scaled problems," according to Annuniziata.
He hesitated to call Quantum System One a brand name like IBM's Watson deep-learning supercomputer platform, but rather a family of quantum systems in the early stages of development.
Watson is a distinct busines unit within IBM, and includes a division called IBM Watson Health. The technologies for the Discovery Accelerator are coming chiefly from IBM Research, though AI will be a major component of the Cleveland center.
In a sense, quantum is a hardware concept, while Watson is more about software.
Annuniziata called System One the "R&D phase of system build" for the long-term goal of revolutionizing computing, noting that the Cleveland Clinic partnership agreement is for 10 years. The planned second generation of the quantum installation in Cleveland will herald what he described as the "practical era" when quantum computers will be able to handle problems large enough to have a direct impact on biomedical research and clinical practice.
This collaboration is about "bringing quantum computation to real problems, starting with small-scale versions of those problems and understanding what the approach would be and how it would scale," Annuniziata said.
"The way we see quantum computing is more of the future look rather than immediate applications," Jehi added.
She said that current computing technology is "choking" with all the data bioinformatics is generating today. Plus, according to Jehi, the binary construction is somewhat incompatible with life sciences.
"The human body is nothing like zeros and ones. Nature is not zeros and ones," Jehi said. "When it comes to analyzing massive amounts of multiomics data, … imaging, and clinical data, we've been really struggling with doing all of that with the existing compute capabilities of what the binary technology can offer."
According to Jehi, processing sequencing and multiomics data then linking it back to clinical information often overwhelms standard binary computing. The same goes for in silico simulation of molecules that potentially could be developed into drug compounds.
"It takes too long, and some questions just cannot be answered," she said. "Quantum could help by accelerating these analyses."
The Cleveland Clinic wanted to get in front of quantum technology to burnish its image as a leader in healthcare delivery and research, according to Jehi, plus she said that the institution likes to experiment.
"We are going into this with eyes wide open that this is an experiment. It may work. It may not work. There will be ups and downs to it," she said.
"We don't claim to know what the applications of quantum are in biomedical science, but we want to discover them," Jehi added. "That's why it was important for us to have the quantum in Cleveland so that we can have the team with it and not just have access to quantum remotely."
The Global Center for Pathogen Research & Human Health will add 400,000 square feet of research space to the half-million square feet the Cleveland Clinic already has for its 200 labs and 1,500 investigators. The Discovery Accelerator and Quantum System One will eventually be housed in the new space, though groundbreaking for the Pathogen Center is not scheduled until spring 2022.
In the meantime, the Discovery Accelerator will begin its work with quantum technology IBM has built in the cloud, as well as other high-performance computing technologies from existing partnerships.
Before Quantum System One is realized, IBM and the Cleveland Clinic plan to deploy a locally hosted, less-powerful quantum installation in existing clinic space by the middle of next year.
Annuniziata said that applying quantum computing to biomedical research is the longest-term project of the Discovery Accelerator. "Codeveloping [this] technology is going to help accelerate discovery and treatment in the pathogen space," he said.
Jehi said that the Pathogen Center also has researchers working on cancer, immune system diseases, and their impact on population health. Although Ohio made the investment, Cleveland Clinic decided to call it a global center because the institution has facilities in Florida and Abu Dhabi, plus one slated to open in London later this year, as well as affiliations with health systems across the US.
The partners are trying to build an ecosystem in Cleveland for life sciences researchers worldwide. "We're very open," Annuniziata said.
Reaching that goal is a two-part effort. Filling the promised 7,500 jobs will take training in new skills including quantum computing, which is where an agency like JobsOhio comes in.
The other part is looking for additional collaborators to further the mission of the Discovery Accelerator and the Pathogen Center.
"We're going to develop a program together which will certainly reach beyond just Cleveland Clinic and IBM and hopefully create that talent and drive that talent in Ohio," Annuniziata said.
Jehi said that the clinic already has several current projects in mind for the Discovery Accelerator, including work on predictive modeling for progression and severity of COVID-19 cases. A Cleveland Clinic team is working on new COVID-19 vaccine nanotechnology, while another is studying reinfection risk.
Some of the clinic's pathogen research is underway at the organization's Florida Research & Innovation Center, including for tropical diseases like Zika.
Jehi said that in silico work like predictive modeling does not require a wet lab, so projects can move to cloud-based quantum computing right away.
Jehi, a longtime Cleveland Clinic neurologist, became CRIO in January 2020. Within a few weeks, she got involved in a collaboration with IBM to search for COVID-19 drug repurposing candidates, even before the novel disease was declared a pandemic.
Using public COVID-19 data, the clinic gained access to IBM supercomputing power and was able to identify several candidates in a couple of weeks rather than months, Jehi said. Preliminary work was published in PLOS Biology in November, and two compounds, melatonin and toremifene, are now in a clinical trial as potential treatments for reducing the severity of COVID-19 infections and preventing hospitalizations.
Given the likelihood that COVID-19 will become endemic, future generations of vaccines remain in development. Jehi expressed hope that the Pathogen Center can address both the long-term consequences of the current pandemic, but prepare for future outbreaks of novel diseases so public health officials and researchers can move fast at the first sign of a problem.