Basic Areas of Research
Proteins already solve a vast array of technical challenges: in nature, they mediate the use of solar energy to manufacture complex molecules, respond to small molecules and light, convert chemical gradients to chemical bonds and transform chemical energy into work — just to name a few.
Our researchers draw inspiration from nature’s proteins and work to design equally useful molecules from scratch.
Protein Structure Prediction
We are using deep learning to quickly and accurately model the three-dimensional protein structures from sequence data alone. Our latest tool, RoseTTAFold, is available for public use.
New Protein Scaffolds
In the past, almost all protein design efforts have modified naturally occurring protein backbones. However, for most challenges, there is unlikely to already exist a protein with an optimal 3D structure. We are developing methods for designing a wide range of exceptionally stable protein structures with tunable geometries for specific applications.
Protein and Small Molecule Binding
We are developing methods for designing high-affinity protein binding and applying these methods to create binders to targets of medical interest. These efforts are providing fundamental insights into the protein-protein interactions which underlie most cellular processes. We are also developing methods for designing proteins that bind with high affinity to small molecules and applying these methods to design binders for drugs with narrow therapeutic windows, toxic compounds and other small molecules of interest. These efforts inform our understanding of small molecule recognition in biology.
Self-Assembling Nanomaterials and Vaccines
Self-assembling protein materials play critical roles in biology. IPD researchers are developing new self-assembling nanostructures and using these approaches to develop the next generation of vaccines and drug delivery vehicles.
Enzymes catalyze chemical reactions that are essential for life. We are developing general methods for creating catalysts for chemical reactions not catalyzed by naturally occurring enzymes.
Forcefield Development & Sampling Algorithms
Our protein design methods seek the lowest energy amino acid sequence given constraints specifying the problem of interest. The more accurate the forcefield used to calculate energies, the higher the activity and success rates of the designed proteins. We use a combination of physical chemistry and analysis of over one hundred thousand protein crystal structures to improve our description of protein energetics on the atomic scale.
Parallel Synthesis and Screening
Once low-energy designed proteins have been identified on the computer, it is critical to test them experimentally. Since neither the design tools nor the sampling methods are perfect, we experimentally manufacture and measure the activity of as many designs as possible. IPD researchers are developing methods for testing tens of thousands of different computational designs in parallel.
Protein Structure Determination
Our researchers develop methods for solving protein structures using limited experimental data that in use in laboratories around the world.