GENESIS: Bridging High-Performance Computing and Cloud for Accessible Biomolecular Modeling
The Department of Defense requires quick and immediate access to new digital
tools to produce novel biotechnologies; However, advanced open source and
commercial tools such as AlphaFold3 require extensive setup and powerful resources,
posing a challenge for DoD scientists needing quick access. Cloud environments can
be costly, yet dedicating personnel to maintain low-cost HPC pipelines diverts scientists
from critical research.
Here we present GENErating Structures In Silico (GENESIS) an optimal solution
that answers unlocks the latest in AI/ML tools for scientists. GENESIS is an innovative
web platform designed to facilitate access to biomolecular modeling predictors simply
by providing protein/RNA/DNA/ligand sequences. GENESIS addresses a critical need
for user-friendly tools, making advanced computational biology accessible to
researchers and enthusiasts without requiring extensive programming or HPC
knowledge.
A core focus of GENESIS is its unique hybrid infrastructure, seamlessly
integrating HPC systems with Air Force Google Cloud (AFGC). AFGC hosts the intuitive
frontend website, while HPC handles intensive computation. Our approach balances
efficient resource usage, cost management, and user accessibility. By integrating this
workflow into new and emerging tools, DoD will stay up-to-date with the state-of-the-art
with minimum downtime for integration. A computer scientist/engineer is needed for
integration development, but allows users to self-submit without needing a middle man.
By removing the technical barriers to accessing powerful computational tools,
GENESIS empowers scientists to focus on their primary research objectives, fostering a
more agile and efficient scientific community. This platform not only enhances individual
productivity but also promotes interdisciplinary collaboration by providing a common,
accessible interface for diverse scientific teams. The continued development of
GENESIS will focus on expanding its suite of integrated tools and further streamlining
the user experience, ensuring that the DoD remains at the forefront of scientific and
technological advancement.

Distribution A. Approved for public release: distribution unlimited. AFRL-2025-3190
IMPACT
Accomplishment: Successfully integrated AFRL's approved Google Cloud platform with the High Performance Computing system for biotechnology applications; Result: Greatly increased the accessibility and usability of advanced AI protein modeling saving upwards of 50K USD / month in cloud computational costs and enabling advanced biomaterials development to enhance warfighter survivability and lethality in the future fight.
PRESENTER
Perminov, Alexander
alexander.perminov.ctr@us.af.mil
937-478-0489
AFRL
CO-AUTHOR(S)
Perminov, Alexander
alexander.perminov.ctr@us.af.mil
Stamps, Blake
blake.stamps.1@us.af.mil
CATEGORY
Hybrid Computing (HPC-Cloud)
SECONDARY CATEGORY
AI/ML for HPC
SYSTEM(S) USED
Raider (raider.afrl.hpc.mil)