Bridging the Gap Between Prediction and Reality in Structural Biology

/Abstract My research is about creating the tools, insights and capabilities needed by the field of structural biology to understand structure-function relationships, including catalysis, enzyme-substrate interactions, and structure- based drug design. In the next five years this field will face new challenges as accurate structure prediction generates a firehose of testable hypotheses, shifting the focus of experimental efforts to the “last Angstrom” between predicted and actual atomic positions. This short distance is where the puzzle pieces of biology and chemistry fit snugly together, and it is the details on this scale that are vital to understanding the structure- function relationship. The experiments of the near future must reveal accurate, high-resolution, and damage- free 3D images. Remarkably, the data used to train current prediction models already contain yet-to-be- unlocked information. Single-electron changes can be visible in Macromolecular Crystallography (MX) data at resolutions as low as 3.1 Å, but structural models must be far more accurate to reveal such features. To enable this robust interpretation of experimental data, I will develop simulation-based models that are minimally restrained but still faithfully reproduce observed average-image and diffuse-scatter data. These improved, multi-conformational models will enhance understanding of how macromolecules transmit forces through their interior and how they influence and interact with other molecules. Deducing these intramolecular communication networks requires solving the topological problem of simultaneously satisfying observed density and prior knowledge of chemical geometry. Solving this problem will be made easier by comparing related structures. This will arise from current technologies for correction of non-isomorphism in real space, which I will migrate into reciprocal space, enabling merging of incomplete data such as XFEL stills and parametric structure frameworks. These low-dimensional frameworks will allow selection from a continuum of 3D molecular structures like a marionette by dialing in desired parameter values, such as temperature, pH, or other reaction progressions. I will test these framework models against the thousands of non-isomorphous data sets collected at my beamline and report on best practice. Radiation damage will be the final barrier, so to move towards damage-free data from a synchrotron, we will implement a new kind of data collection called "painting with X-rays," leveraging modern fast-framing detectors and small X-ray beams to extract the full information content from each sample and extend zero-dose extrapolation to the single-photon level. Collectively, I expect the benefits of bridging the gap between prediction and reality to be transformative to both methods development and functional studies using complementary structural techniques, such as CryoEM, SAXS, tomography electron diffraction, and especially hybrid methods that combine structural data from multiple sources. Project Number: 1R35GM158447-01 | Fiscal Year: 2025 | NIH Institute/Center: National Institute of General Medical Sciences (NIGMS) | Principal Investigator: James Holton | Institution: UNIVERSITY OF CALIFORNIA, SAN FRANCISCO, SAN FRANCISCO, CA | Award Amount: $501,000 | Activity Code: R35 | Study Section: Maximizing Investigators' Research Award B Study Section[MRAB] View on NIH RePORTER: https://reporter.nih.gov/project-details/11090132