Latest News

Computing transforms chemistry, biotech


Reprint of article originally published March 19, 2016 in The San Diego Union-Tribune 

By Bradley J. Fikes

The life sciences are going dry.

While test tubes, Petri dishes and cell cultures abound in the biotech world, the frontiers of exploration are increasingly taking place on high-powered computers and supercomputers.

Equipped with ever better microprocessors to perform calculations and render graphics, computers today run extraordinarily powerful software that sheds light on biomedicine at every scale — from the physics of atomic-level interactions to sophisticated models of how tissues function.

The American Chemical Society spotlighted this trend during its just-concluded conference in San Diego, with its theme of computers in chemistry.

Computers don’t replace biotech’s “wet” chemistry, but make it more productive. Three decades ago, computers served as tools in a few then-exotic areas such as atomic-level drug design. Today, they are mainstays throughout the entire process of developing medications.

Software programs contain vast amounts of biological data, helping researchers decide on a good drug target. They also predict a test medication’s properties, record experiment information and store data from clinical trials.

In the process, chemists have found new opportunities to prove their worth.

Presentations at last week’s convention were packed with examples of chemistry and supercomputing, such as:

--The San Diego-based biotech company Actavalon used computers to identify candidate drugs that, if successful, would battle cancer by reactivating an important tumor-suppressing gene called p53.

--New York City-based company Schrödinger has developed software to understand how mutations affect the characteristics of proteins, the workhorse molecules of the body.

--Scientists at Dartmouth College in New Hampshire used computers to reduce the body’s adverse reactions to an enzyme that attacks the deadly bacterium MRSA. The company Stealth Biologics is harnessing this technology for drug development.

“Computing is transforming how we do chemistry,” said Rommie Amaro, a computational chemist at UC San Diego and co-founder of Actavalon who gave a Kavli Foundation lecture on the topic at the American Chemical Society meeting.

In the early 1990s, supercomputers could model systems as complex as a single protein made up of about 10,000 atoms, Amaro said. Today, they can model entire viruses, which contain hundreds of millions of atoms.

Amaro is also a mentor to Eric Chen, the San Diegan who in 2013, at age 17, won the grand prize from the Google Science Fair and other major science competitions for high school students.

At the time, Chen was working under special arrangement with Amaro and other UC San Diego researchers. He used the university’s supercomputer to screen 450,000 molecules for activity against a key influenza enzyme. The screening produced 237 hits against that enzyme, which laboratory testing then winnowed down to six.

As for Chen, he enrolled at Harvard University in 2014. He is now also a research scientist at the Massachusetts Institute of Technology.

Chemistry-biotech fit

Chemistry studies the structure, composition and properties of matter.

In practice, this often means the study of chemical reactions. For example, the addition of table salt to water produces a substance similar to water, but with additional properties derived from the dissolved salt.

Biotechnology uses living things or life processes to make products. For instance, in one of the field’s first applications, inserting a gene for human insulin into bacteria allows them to produce human insulin. More recently, genetically re-engineering the immune cells of blood cancer patients directs the cells to attack the cancer.

These fields seem rather distant from each other, but all genes are made up of four chemicals strung together in a molecule called DNA. These “letters” of the genetic alphabet have distinct properties determined by their molecular structure. If you’re going to tinker with DNA, you’ve got to understand chemistry.

Proteins, too, can’t be understood without chemistry.

These large molecules are made out of amino acids, simple molecules that DNA codes for. There are 21 amino acids naturally made by living things, plus scores of “unnatural” amino acids made in the lab. Adding these synthesized amino acids to natural proteins can give those proteins new properties. How unnatural amino acids change the activity of proteins is a matter of chemistry as well.

Computers began making a big mark in the 1980s, with what was then called “rational drug design.” The goal was to use computers to help understand a disease process at an atomic level — the building blocks of amino acids and proteins — and then design a drug to stop that process.

That contrasts with the hit-or-miss approach of throwing many compounds at a target and looking to see what, if anything, works.

San Diego’s Agouron Pharmaceuticals, founded in 1984, became a success story with its rational design of an anti-HIV drug called Viracept. The medication and Agouron’s overall potential led to its 1999 purchase by drug giant Warner-Lambert for $2.1 billion. The next year, Pfizer merged with Warner-Lambert in a $90 billion merger.

As a result of the second merger, Pfizer acquired operations in La Jolla that it still holds today.

Rising power

Rational drug design, now usually just called drug design, continues as an important way to take advantage of novel biological information, such as locating a new vulnerable spot on HIV or the Ebola virus.

In the 1990s, computers made themselves useful in a huge initiative that still resounds today — the Human Genome Project.

Sequencing all 3 billion letters of the human genetic alphabet and confirming their accuracy proved to be difficult and expensive, yet doable. But to move from one sequence to understanding all the mutations and their interactions in billions of people demands the highest-performing computers to analyze and store vast amounts of data.

Companies led by San Diego’s Illumina are making this possible.

While that large-scale work proceeds, smaller biotech firms such as San Diego’s Actavalon are relying on computers to solve narrower but important problems such as reactivating the cancer-suppressing p53 gene.

Actavalon was founded in 2013 to find therapies for p53-related cancers, and mutated p53 is found in most cancers.

Toward that end, Amaro (the UC San Diego professor and Actavalon co-founder) received a $200,000 grant from the NVIDIA Foundation, the philanthropy set up by the eponymous computer graphics card manufacturer. Awarded in 2013, the grant was used to develop software that takes advantage of the computing power of graphics processing units in graphics cards.

As it turns out, the ability of graphics cards to render realistic game imagery also vastly speeds up the ability to simulate the motions of molecules.

Amaro wanted to find regions of p53 that can attach to drugs, potentially restoring the protein’s proper activity. Her team found these regions with molecular simulation software. The regions weren’t visible with the traditional visualization method of X-ray crystallography, used to discover the structure of DNA.

X-ray crystallography generates static images, but computer simulations show the actual range of atomic-level vibrations and changes in the shapes of molecules, Amaro said in her American Chemistry Society presentation.

These extremely rapid movements can open up parts of molecules to bind with others. Understanding these configurations is necessary to predict which drugs will interact with which proteins.

If a protein folds into a certain shape creating a concave pocket, a drug that fits into that pocket could enter and attach to the protein — like a key in a keyhole. To extend the analogy further, the keyhole may only open briefly then disappear. Computer simulations can show when that keyhole appears, and for how long.

“Nowadays, these (graphics) chips have become supercomputers on the desktop,” Amaro said.

Link to the full article from the San Diego Union-Tribune

Fikes, Bradley J. "Computing Transforms Chemistry, Biotech." San Diego Union-Tribune. San Diego Union-Tribune, 19 Mar. 2016. Web. 01 Feb. 2017.