||Dr. Arthur Sherman
For the layperson, he translates: “Math uses equations to describe how systems evolve in time, whether the systems are physical, biological or economic. If you know something’s current state and the laws of how it evolves, you can track how it moves.” That’s what gives math its predictive power, as in Newton’s laws of motion.
To Karen Ong, math is cool. Now entering her second year as a post-baccalaureate IRTA fellow, she has created a presentation for LBM’s 50-year anniversary. From its origins in 1957, she says, the lab linked mathematicians and biologists
to collaborate in groundbreaking ways.
“This is what is magical about what we do,” says Ong. “We take reality, abstract it into math, construct a model and manipulate it.”
And, she says, the work is dynamic: “You can get data from experiments and go back to math. You can start with theory and go to the experimental model. Then you verify, or modify the model.”
Such research can mathematically do many more experiments in the time it takes to do a single “real” experiment. For example, screening
molecules computationally, rather than in a wet lab, can reduce the time needed to find promising drug candidates. Doing experiments
“in silico”—on a computer or by computer
simulation—can also shape the direction
and interpretation of experiments in vitro and in vivo.
Mathematical biologist Joel E. Cohen of Rockefeller
and Columbia universities has described how we depend on math for understanding cells and their signaling; the brain, behavior and emotion;
genes, genomes and prions; the biospheres and global processes; epidemics and ecology; and how these systems interact. Math, like any living
language, not only adapts to the challenge of what it’s describing, but also gives insight. Mathemetizing
inspires conceptual leaps.
“It’s like a piece of art,” Ong says. “You can’t model every detail, but you can paint the key aspects and from that picture make statements about the original system.
|Karen Ong, IRTA fellow, created a presentation documenting LBM’s 50-year history.
“Moreover,” she adds, “this [biological] system is a lot more complicated than what you can figure
out intuitively. Using math, you can find out things you wouldn’t have thought to look for.”
These things include a spectrum of interests. “The main application for me,” says Sherman, “has been the biophysical basis of insulin secretion
in pancreatic beta-cells.” He joined the lab as a postdoc in 1986 and became chief in 1997.
LBM’s four principal investigators and one research scientist, plus (depending on the academic
cycle) around 12 postdoc and 2 postbac fellows, are working on a variety of theoretical and computational approaches to diabetes, obesity
and neuroscience, including mitochondrial function and insulin action; insulin resistance and food intake regulation; obesity clinical trials;
and structure-based drug design. Investigations
range from extremely applied (drug discovery and clinical data from subjects in the metabolic lab) to extremely theoretical.
One of the current principal investigators, Dr. Carson Chow, stresses the foresight of Dr. Frederick
S. Brackett, a member of NIAMD’s Laboratory
of Physical Biology. Early on, Brackett understood how important computers would become [see sidebar].
“While the current NIH Roadmap is emphasizing systems and computational biology,” Chow says, “I think very few people know that the LBM has actually been in the game for 50 years. Wilfrid Rall, one of the pioneers of computational
neuroscience and cable theory for dendrites, was an early member of the lab. Other notables include John Hearon and Mones Berman, both of whom were instrumental in developing the theory of compartmental modeling and tracer kinetics. Much of the mathematics of modeling
electrical activity in neurons and endocrine cells was developed [here], especially when John Rinzel was chief.” He adds that scientific director
Dr. Marvin Gershengorn, a strong supporter, added PIs to “build the lab up.”
In 1957, Brackett chaired a committee that recommended forming a Mathematics Panel. As this became a branch in its own right—the Mathematical Research Branch (MRB), now known as LBM—it could explore theoretical biology without being limited to consulting as a service to NIH.
|NIH has not always been so enthusiastic about using computers in research. In the late 1950s, Dr. Frederick S. Brackett and other members of the electronic
data processing committee faced overwhelming opposition to their proposal
to acquire an IBM 650 for scientific purposes. Then-NIH director Dr. James Shannon approved the EDP committee’s plan in January 1957, but the division chiefs banded together to prevent its realization. They argued that general-purpose
digital computers like the 650 could not be kept sufficiently busy by NIH researchers who, they believed, lacked the knowledge of mathematics and electronics
to gainfully employ computers.
To overcome such doubts, Brackett organized the Mathematics Panel to prepare
researchers for computer use by helping them mathematize their work. Nevertheless, funding for computer work remained scarce until 1960, when Shannon was able to circumvent the objections of the division chiefs by acquiring
money for biomedical computing directly from the U.S. Senate, then concerned
that the United States was trailing the U.S.S.R. scientifically.
Biomedical computing’s most vocal patrons were Senators Hubert H. Humphrey
(D-MN) and Alexander Wiley (R-WI). Between 1956 and 1959, NIH spent less than $2 million on computers in research (intra- and extramurally),
but for the period between 1960 and 1963, that figure jumped to over $40 million.
“We now have collaborations with other ICs and outside NIH,” Sherman says. “NIDDK produces
ligands, guided by calculation or computer
screening, or drawing from a large library of 100,000 compounds…and we use computers to accelerate drug discovery.”
A ligand is an atom, ion or molecule that shares its electrons with another (central) ion. Just as ships dock at a pier, ligands are said to “dock” on receptors in a cell.
Drug discovery became a component of the lab only recently. Much of the history of LBM was involved in developing and analyzing mathematical
models of various physiological systems such as neurons, the pancreas, mitochondria, human metabolism, cortical circuits, etc.
As for how LBM came to be housed in NIDDK: “It’s a historical accident, a historical development,”
Sherman explains. “It was originally suggested to be within the OD because of the expected cross-institute nature of the mathematical
research, but since no research was permitted in the OD, NIDDK—NIAMD at the time—agreed to take it on for administrative convenience.”
He and Chow stress that other ICs also have mathematical modeling labs: NICHD’s Laboratory
of Integrative and Medical Biophysics; NIAID’s Program in Systems Immunology and Infectious Disease Modeling; CIT’s Mathematical
and Statistical Computing Laboratory; and “of course,” Chow adds, “the NCBI [NLM’s National Center for Biotechnology Information]
is partly a mathematical biology institute.
It’s just that [our group] has been around for 50 years and most people don’t know that.”
And the genome connection?
“Other ICs have computational groups,” Sherman says. “But we gave birth to NCBI.” NCBI director Dr. David Lipman spent postdoc time in the MRB, where, Sherman adds, ideas for bioinformatics and the NCBI took flight.
“When the genome project took off,” he says, “it outgrew the lab here. It was time for the cuckoo to leave the nest, and now it dwarfs its parents.”
As for the future, “I’m not good at prognostication...my bias is that answering
big questions will come out of answering small questions, and patterns will emerge. The place where there has been success in math models has been traditional
“We have a long way to go to understand how modules work,” he continues. “Maybe they have grand unifying principles. But biology does what it does because it pyramids a thousand different components together. We’re still at the stage of understanding it, component by component.”