Drugs work by binding to a specific protein molecule and changing
its three-dimensional shape, altering the way it works once inside the body. The ideal
drug is designed in a shape that will only bind to a specific protein or
proteins involved in a disease while eliminating side effects that
occur when drugs bind to other proteins in the body.
Designing successful drugs is like solving a puzzle. Without knowing the three-dimensional shape of a protein, it would be like trying to solve that puzzle with a blindfold on.
‘A new set of machine learning algorithms developed by researchers can generate 3D structures of tiny protein molecules. This may revolutionize the development of drug therapies.’
A new set of machine learning algorithms developed by U of T
researchers that can generate 3D structures of tiny protein molecules
may revolutionize the development of drug therapies for a range of
diseases, from Alzheimer's to cancer.
U of T PhD student Ali Punjani, who helped develop the algorithms, notes, "The ability to determine the 3D atomic structure of protein
molecules is critical in understanding how they work and how they will
respond to drug therapies."
This new set of algorithms reconstructs 3D structures of protein
molecules using microscopic images. Since proteins are tiny - even
smaller than a wavelength of light - they can't be seen directly without
using sophisticated techniques like electron cryomicroscopy (cryo-EM).
This new method is revolutionizing the way scientists can discover 3D
protein structures, allowing the study of many proteins that simply
could not be studied in the past.
Cryo-EM is unique because it uses high-power microscopes to take
tens of thousands of low-resolution images of a frozen protein sample
from different positions. The computational problem is to then piece
together the correct high-resolution 3D structure from the
low-resolution 2D images.
"Our approach solves some of the major problems in terms of speed
and number of structures you can determine," says Professor David Fleet,
chair of the Computer and Mathematical Sciences Department at U of T
Scarborough and Punjani's PhD supervisor.
The algorithms, which were co-developed by Fleet's former
Post-Doctoral Fellow Marcus Brubaker, now an Assistant Professor at York
University, could significantly aid in the development of new drugs
because they provide a faster, more efficient means at arriving at the
correct protein structure.
"Existing techniques take several days or even weeks to generate a
3D structure on a cluster of computers," says Brubaker. "Our approach
can make it possible in minutes on a single computer."
Punjani adds that existing techniques often generate incorrect
structures unless the user provides an accurate guess of the molecule
being studied. What's novel about their approach is that it eliminates
the need for prior knowledge about the protein molecule being studied.
"We hope this will allow discoveries to happen at a ground-breaking
pace in structural biology," says Punjani. "The ultimate goal is that it
will directly lead to new drug candidates for diseases, and a much
deeper understanding of how life works at the atomic level."