Purdue University

The Southeastern Center for Microscopy of Macromolecular Machines (SECM4) utilized Purdue University’s Anvil supercomputer to host the 2024 SECM4 data processing workshop. The workshop focused on teaching researchers how to process data for single-particle cryogenic electron microscopy (SPA-cryo-EM) analysis.

Dr. Nebojša (Nash) Bogdanović is a faculty member specializing in cryo-EM who co-manages the operations of the SECM4 cryo-EM service center, located at Florida State University. SECM4 is a National Institutes of Health-funded resource aiming to democratize high-end, often high-cost cryo-EM instrumentation for underrepresented research groups. It supports broader free-of-cost access across the Southeastern U.S. and other IDeA states, with much of its effort dedicated to training biomolecular researchers in the region. Dr. Bogdanović’s role at SECM4 includes designing workflows and optimizing approaches for high-resolution structure determination. He trains users in fundamental and advanced cryo-EM strategies, including vitrification of very large numbers of diverse samples, microscope operation, efficient data collection, and workflow management with data processing. To further the advancement of cryo-EM research and analysis, Dr. Bogdanović developed the materials for and taught the 2024 SECM4 data processing workshop, where attendees learned how to use HPC-based cryo-EM software such as Relion, CryoSPARC, and ML-based Topaz for tasks like preprocessing, particle picking, 2D and 3D reconstruction, classification, and model building. This workshop was the second in a row of hands-on training events.

Cryo-EM is a form of transmission electron microscopy in which delicate samples, such as proteins, viruses, DNA, RNA, or cellular structures, are cryogenically frozen into vitreous (glassy) ice before being observed and imaged. Rapidly freezing samples in this manner prevents ice crystal formation and enables higher-resolution reconstructions. Thousands of micrographs, or “movies,” are captured and processed through specialized, computationally demanding workflows to generate sufficient structural data for reliable 3D models. These models offer researchers unprecedented atomic-level detail indispensable for understanding biological mechanisms and designing highly specific drugs. Although cryo-EM is revolutionizing structural biology, single-particle cryo-EM analysis is both complex and data-intensive. Researchers must be properly trained in software usage, the step-by-step analysis process, and strategies for handling large datasets. Recognizing this need led to the creation of the SECM4 cryo-EM workshops as well as a very popular, dedicated cryo-EM consultation service accessible to SECM4-approved projects.

The 2024 SECM4 data processing workshop was a big success for a growing center. It was attended by 12 researchers from internal and external institutions interested in learning more about cryo-EM workflows applicable to their projects. Dr. Bogdanović led the course alongside Dr. Scott Stagg, the acting Facility Administrator for SECM4. Due to the nature of cryo-EM work and the workshop’s size, the instructors required a resource capable of providing large-scale computing power. They turned to Anvil for that support.

“We understood that the computational resources required for this work are very intense,” says Dr. Bogdanović. “So we needed 4 to 8 GPUs, 500 GB to 1 TB of RAM (or more), as well as a large SSD allocation. What we managed to do with Anvil was to use their implementation of CryoSPARC, a software we use readily in our field, and distribute it to the 12 participants in our workshop to demonstrate how each step is carried out.”

The Anvil team provided the SECM4 workshop with access to the supercomputer’s advanced GPUs and granted a special dispensation to reserve a block of GPUs for the three-day course.

“So we, thanks to the kindness of the Anvil team, were able to reserve up to 10 GPUs simultaneously,” adds Dr. Bogdanović, “guaranteeing that our participants could run jobs during the workshop. That worked out wonderfully, and we are very grateful Anvil was able to do this.”

Dr. Bogdanović was delighted with Anvil’s performance. To prepare for the workshop, he gained access to Anvil months in advance through a proposal-based, NSF-funded ACCESS program, ensuring the system would fit his needs. CryoSPARC was already installed and ran flawlessly, better than his experience with the software on other HPC systems. RELION was also available, but he needed a different version for the workshop. The Anvil team was on hand to help and guided him through their specifics of installation, and RELION worked perfectly when implemented. Dr. Bogdanović prepared all the results for the workshop projects in advance to determine what was most suitable for inclusion and to keep backups on hand in case of any hiccups. Fortunately, everything ran smoothly, and the workshop was a huge success. Participants found it so useful that five went on to apply for and receive their own Anvil allocations. Overall, Dr. Bogdanović was thrilled with how the workshop turned out and hopes to host another later this year.

To learn more about the 2024 SECM4 data processing workshop, please visit: https://www.secm4.org/2024-processing-workshop

For more information about HPC and how it can help you, please visit our “Why HPC?” page.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program, which serves tens of thousands of researchers across the United States.

Researchers may request access to Anvil via the ACCESS allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu