Skip to main content
Premium Trial:

Request an Annual Quote

QuantumBio Says its Algorithm Could be Quantum Leap for Virtual Screening


Virtual screening relies on precise calculations of binding properties between a small molecule and an active site, but most current methods involve performing these calculations for only a few hundred atoms, rather than the tens of thousands of atoms that are involved in binding. This limits the accuracy of the final result, but it’s a necessary tradeoff because atomic-level calculations for the entire molecule would be so computationally demanding that the task could not be completed in a reasonable amount of time.

Pennsylvania State University spin-off QuantumBio, however, claims it has found a way around this limitation. As its name suggests, the company is applying the atomic-scale methodology of quantum mechanics to biomolecular modeling — a task that would be intractable without a proprietary divide-and-conquer algorithm it developed to reduce the computational expense of the process.

QuantumBio founder and CEO, Ken Merz, developed the algorithm while at Penn State, where he still holds a full-time position in the department of chemistry. Merz said that the approach scales linearly with the number of atoms, instead of exponentially — a speed-up that enables calculations for 10,000 atoms in a single day on an ordinary laptop.

QuantumBio is commercializing the Merz algorithm as a computational software suite called DivCon Discovery Suite, which contains a scoring function suite called QMSCore. “Scoring is a pain point for everyone” doing virtual screening, Merz said, but most current approaches are “very highly simplified” for performance purposes, and rely on classical mechanics, electrostatics, and hydrophobicity to calculate binding probabilities. “All these things are trying to capture basically the quantum reality of an interaction of the small molecule with the pocket,” Merz said.

The inability of current methods to provide atomic-scale granularity can have a serious impact on accuracy. “Say you’re taking about a 10 kcal/mol binding of free energy for a good inhibitor to a protein target,” said Merz. “If your force field is off by one kcal per interaction, and there’s five interactions, you may predict that to be a 5 kcal/mol binder and then discard it and never consider it; or you may predict it to be 15 kcal, and then you would run off, synthesize it, and think you solved all the world’s problems with this particular molecule for this disease state, and it turns out that it’s not as good as you thought it [was].”

In the future, the company plans to offer its own docking algorithm that is “optimized to benefit our scoring,” Merz said, but DivCon currently works with any docking algorithm. The DivCon Discovery Suite, which contains a total of six modules, is available as a standalone package, or as part of the company's broader Chemix platform. QuantumBio began shipping version 1.0.0 of Chemix, and version 2.1.0 of the DivCon Discovery Suite, in mid-May. Chemix comes with another product, called HAMStER (Haptic Application for Molecular Structure and Energy Refinement), which adds force-feedback capability to the visual interface so that researchers can “feel” as well as see how well a molecule fits within a pocket.

Merz said that QuantumBio has several customers for its software, and that a few other companies are using it on a trial basis, but he did not disclose the identity of these firms. On the partnership front, the company recently became “an optimized software partner” for IBM Life Science and Healthcare, said Andrew Chu, vice president of business development for QuantumBio.

The company was founded in 2002, and has raised around $1.1 million in funding so far from “friends and family,” the Benjamin Franklin Technology Partners fund, and the Pennsylvania Life Sciences Greenhouse.

The Merz lab at Penn State is closely involved in the company’s R&D efforts, and academic code from the group eventually makes its way to the company’s software development unit — a team of about four developers led by Lance Westerhoff, chief software engineer and product manager — where it is refashioned into user-friendly commercial form.

“We can’t expect grad students to check out GUI code that they’re never going to publish,” said Westerhoff. “So the idea is that as much of that non-technical stuff [as we can], we’ll try to do in the development group, and then in the Merz [lab] they do more of the science.”

Merz said there are about 15 people in his lab at the university, while QuantumBio employs about 10 people. The close ties with Penn State also serve a fiduciary purpose, he said, by providing him with a steady paycheck as a full-time professor while the company works through its “bootstrapping” phase.

But Merz and his colleagues are confident that their approach will catch on. “If quantum mechanics was just as [computationally] expensive as classical mechanics, no one would be using classical mechanics,” Merz said. “In my opinion, the only thing that’s holding people back is software that’s easy to use.”

— BT


Filed under

The Scan

Expanded Genetic Testing Uncovers Hereditary Cancer Risk in Significant Subset of Cancer Patients

In Genome Medicine, researchers found pathogenic or likely pathogenic hereditary cancer risk variants in close to 17 percent of the 17,523 patients profiled with expanded germline genetic testing.

Mitochondrial Replacement Therapy Embryos Appear Largely Normal in Single-Cell 'Omics Analyses

Embryos produced with spindle transfer-based mitochondrial replacement had delayed demethylation, but typical aneuploidy and transcriptome features in a PLOS Biology study.

Cancer Patients Report Quality of Life Benefits for Immune Checkpoint Inhibitors

Immune checkpoint inhibitor immunotherapy was linked in JAMA Network Open to enhanced quality of life compared to other treatment types in cancer patients.

Researchers Compare WGS, Exome Sequencing-Based Mendelian Disease Diagnosis

Investigators find a diagnostic edge for whole-genome sequencing, while highlighting the cost advantages and improving diagnostic rate of exome sequencing in EJHG.