In recent years, the field of cryo-electron microscopy (cryoEM) has undergone a major transformation that has resulted in the ability to resolve structural details at levels tha are comparable to X-ray crystallography. Advanced electron microscope optics are now capable of generating extremely stable and coherent high-resolution electron beams, and recent developments in direct detection sensors enable us to image scattered electrons with extremely high sensitivity and accuracy. These advancements are making a significant impact on biomedical sciences, as structure-based approaches play a key role in identifying and characterizing the determinants of disease and in developing therapeutics. The Scripps Research Institute (TSRI), the largest non-profit biomedical research institute in the United States, has established a world-class cryoEM facility, with more high-resolution microscopes and direct electron detectors than any other site in the country. Automated data acquisition and streamlined image processing software, developed at TSRI, enables high-throughput structure determination of macromolecular complexes. Whereas data acquisition was historically the bottleneck in high-resolution structure determination, the cryoEM infrastructure established at TSRI enables rapid collection of the massive datasets that are produced using the current technology. This has therefore led to an enormous increase in the computational requirements necessary to produce atomic resolution structures from these data. As the imaging instrumentation improved, the size and quantity of the images increased dramatically, to the point that structure each determination by cryoEM now requires hundreds of compute nodes equipped with large amounts of memory. Hence, although we are able to acquire high-resolution data at a rate that is unparalleled by any other facility in the country, we are considerably impaired in our ability to process and analyze these data. Without commensurate computational resources to match our rate of data collection, our ability to fully capitalize on ou world-class cryoEM resources remains limited, as is our ability to optimize data collection strategies and processing parameters, which impedes the development of new technology and algorithms to extend the resolution limits of cryoEM structure determination. More importantly, these limitations directly impact our quest for insights into the molecular basis of human disease, and the development of effective treatments. The installation of such large-scale computing infrastructure will greatly facilitate high-throughput atomic resolution structure determination of macromolecular complexes, as well as development of novel computational strategies for data processing. The proposed infrastructure will include appropriate storage devices, and a centralized high- performance computing environment for image analysis and algorithm development. The computational infrastructure proposed here will significantly enhance state-of-the-art cryoEM structure determination and lead to exciting new discoveries in biomedical research.
Knowledge of the three-dimensional shapes of the cellular factors and protein antigens from microbial pathogens that lead to disease significantly improves our ability to design vaccines and therapeutics to specifically target or prevent these diseases. Cryo-electron microscopy is an increasingly powerful technique for atomic-resolution structure determination of complexes associated with human disease. The Scripps Research Institute has established a world-class cryoEM facility, with more high-resolution microscopes and direct electron detectors than any other facility in the country. However, present day EM structural studies are extremely computationally intensive, and extend beyond the capabilities of our current computational infrastructure. This impedes our ability to address a range of biomedically important questions. To fully capitalize on our unmatched EM facility, we propose the development an equally high-caliber computing infrastructure to enable high-speed structure determination. Together, the resulting infrastructure will provide researchers from a variety of backgrounds with the ability to integrate critically important structural information into their biomedical research.
