Recent improvements in computational methods could lead to a much healthier future
Published April 3, 2017
(L to R): MSU scientists Michael Feig, George Mias, Arjun Krishnan and Alex Dickson are using recent advances in computational methods to enrich their research in areas that include biological systems, precision medicine, genomics and pharmaceutical drugs.
Advances in computational methods could lead to a better understanding of human health and the disease process, or the development of new cures for diseases. Four Michigan State University Department of Biochemistry and Molecular Biology (BMB) scientists are hard at work, pushing the boundaries of these methods to enrich their research.
“People have been doing computer simulations since the computer was invented,” said Alex Dickson, a BMB assistant professor with a joint appointment in the MSU Department of Computational Mathematics, Science and Engineering (CMSE). “But there have been tremendous strides even in the past ten years in our ability to capture longer and longer timescale events, bigger and bigger systems, and to describe these molecular systems more accurately with better force fields.”
Dickson’s laboratory uses computational techniques such as molecular dynamics to study how molecules bind to small proteins in the body.
“Binding and unbinding events occur on very long timescales, and conventional molecular dynamics techniques are applicable only to processes that are on a nanosecond to microsecond timescale,” Dickson said. “To see a drug unbind from its molecular receptor can take from several seconds, up to minutes. That’s a huge gap to bridge.”
To bridge that gap, Dickson has developed a novel method—called WExplore—that allows the study of long timescale events using only microsecond scale simulations.
“This method uses many trajectories in parallel that are periodically managed so we can generate these trajectory sets that are very diverse,” Dickson said. “We can then take these trajectories, put them back together, and explore these ‘landscapes of molecular conformations’ that are much broader than what we could get with traditional simulations.
“We’re principally interested in how pharmaceutical drugs bind to their protein targets, and how we can use this information to design the next generation of drug molecules,” Dickson added.
Professor Michael Feig investigates the structure, dynamics, and energetics of biological macromolecules such as proteins or nucleic acids. Ten years ago, his lab was able to study only one protein at a time. Now, they can look at up to thousands of proteins at once.
“If you’re looking at one protein at a time, you’re looking at a very artificial system,” Feig said. “With computational methods, we can see the real biological environment with everything that’s happening at the same time. We’re getting much closer to what biology really is.”
Feig explained that the interplay between experiments and the computer in the past five to ten years has become incredibly productive.
“Once we have models that describe biological cells accurately and realistically,” he said, “we can use these models to design drugs, change a biological system or maybe even cure diseases.”
Work in the lab of Arjun Krishnan, a CMSE assistant professor with a joint appointment in BMB, focuses on how genomics relates to human health and disease. His approach involves using existing large datasets that represent decades of experimental work by hundreds of researchers across the world who have made their data publicly available.
“The only way to leverage this data in some comprehensive way is to use computational methods that can ‘de-bias’ the large amount of data—by using scarce but high-quality prior knowledge and by focusing on specific biological contexts,” Krishnan said. “Such an approach is relevant to human disease as a vehicle toward mechanistic understanding and genetic diagnosis. One way to improve genetic diagnosis (for a specific disease, such as autism, for example) is to predict in a data-driven manner which genes might be involved in the disease.
“We can discover disease-associated genes and variants that have never been studied before,” Krishnan added. “I think this is important to move biology forward.”
Research in the lab of Assistant Professor George Mias focuses on precision medicine. A specific project in that lab consists of monitoring individuals over multiple time points to see if there are changes to these individuals at the molecular level that reflect the changes they’re undergoing in their general health. Through their clinical trials, they collect blood and immune cell samples, from which they may extract information on more than 20,000 genes, 8,000 proteins and 10,000 metabolites. These thousands of measurements cannot be analyzed by hand.
“To study this kind of information—looking at longitudinal data from thousands of components—we need the help of computers,” Mias said. “Surprisingly, as recently as 2012, the technology and the computational capabilities to analyze these types of measurements were not utilized to help us make sense of this kind of large data.
“Precision medicine is a direction more researchers are moving toward, so we’ll be seeing a much heavier use of computational methods,” he added.
“The role of computation today is to make sense of very complicated experiments, and to develop models that can interpret the experimental data better,” Feig said. “We also use it to develop hypotheses to generate ideas to guide experimentalists in what kinds of experiments to do next, and what kinds of questions to ask.”
“Taking the computational approach, which is very unbiased, allows us to venture into the unknown in a very systematic way,” Krishnan said.