Arizona State University & CR8DL | Orchestrating Cloud-Supported Workspaces For A Computational Biochemistry Course At Large Scale

Read the full research paper here.

At CR8DL, our joint proof-of-concept project with Arizona State University (ASU) aimed to deploy a Jupyter notebook-based interface to datacenter resources for a computationally intensive, semester-length biochemistry course project. For undergraduate biochemistry students with limited high-performance computing experience, this straightforward interface provided access to large-scale computations.

The CR8DL resource addressed critical challenges in the course. Students were assigned a semester-long project involving protein structure prediction using the Alphafold software. To accelerate the computations, CR8DL’s cloud environment provided high-performance computing resources, including CPU, RAM, and the latest GPU compute resources. The Jupyter notebook interface allowed students to access these resources with minimal knowledge of computer system architecture or operating systems.

Throughout the semester, students iteratively used the AlphaFold pipeline to generate structures and assess confidence in output structures using quantitative pLDDT scores. As the project progressed, various enhancements were identified and implemented, including the development of the Parafold utility, which further parallelized computations and significantly accelerated the workflow.

The joint effort between CR8DL and ASU enabled students to achieve substantial progress in generating protein backbone structures similar to the target domain. This success was made possible by leveraging CR8DL’s cloud-supported workspaces, accessible and user-friendly for researchers with limited HPC experience.

As CR8DL continues to evolve, efforts are underway to further streamline the user experience in complex coursework, removing barriers and encouraging iterative engagement with results. By democratizing access to supercomputing resources, CR8DL empowers students and researchers to accelerate discovery and drive innovation.