TLDR: The Defense Advanced Research Projects Agency (DARPA) has launched the Network of Optimal Dynamic Energy Signatures (NODES) program, an ambitious AI-driven initiative aimed at predicting the function of virtually every known protein. This program, which officially commenced in July 2025, integrates advanced artificial intelligence with biophysical analysis to accelerate threat assessment, develop countermeasures, and establish an early warning system against pandemics and biowarfare. Arizona State University is a key contributor, with Professor Abhishek Singharoy involved in achieving rapid protein function prediction.
In a significant stride towards global health security, the Defense Advanced Research Projects Agency (DARPA) officially unveiled its Network of Optimal Dynamic Energy Signatures (NODES) program in July 2025. This groundbreaking initiative is set to fundamentally transform humanity’s defense against future pandemics and biowarfare threats by leveraging advanced artificial intelligence to predict the function of virtually every known protein. The program’s immediate implications are profound, promising to accelerate threat assessment, countermeasure development, and establish an unprecedented early warning system for biological dangers.
The NODES program represents a critical strategic pivot towards proactive biodefense, moving beyond reactive responses to emerging pathogens. At its core, NODES is the first biophysics-guided deep learning framework designed to input protein sequences and predict their associated functions based on protein folding, protein binding, and allosteric interactions. This sophisticated approach moves beyond static structure-based models, employing high-throughput molecular dynamics (MD) simulations integrated with AI and machine learning (ML) frameworks to decipher the complex movements and structural changes of proteins.
Currently, out of over 300 million protein sequences and structures cataloged, only about 1% are functionally characterized, leaving approximately 90% of the known protein universe functionally uncharacterized. Traditional models often fail to account for protein motion and context, a gap NODES aims to fill by modeling dynamic behaviors. This capability is crucial for rapid functional annotation, even for previously unknown molecules.
Key scientific contributions to the NODES program include those from institutions like Arizona State University (ASU), where Professor Abhishek Singharoy is on special assignment with DARPA, contributing to the program’s ambitious goal of predicting protein function within an hour. This collaboration underscores the interdisciplinary nature of the program, combining expertise in biophysics, AI, and computational biology.
Also Read:
- ASU Professor Vidya Chhabria Honored with Google Award for Pioneering AI in Chip Design
- New AI Model Predicts Genetic Mutation Pathogenicity by Analyzing Protein Dynamics
The anticipated outcomes of the NODES program are far-reaching. It is expected to culminate in a universal design platform for synthetic biology, disease pathway mapping, and biosecurity applications. Furthermore, NODES will be able to estimate the evolutionary probability of protein sequences, helping to determine whether a molecule has arisen naturally or has been synthetically engineered, thereby aiding in the detection of potential biothreats. The program is structured as a 39-month effort, divided into a 12-month Phase 1 Base Period, a 15-month Phase 2 Option Period, and a 12-month Government-only Phase 3 Option Period, with performers addressing three functional areas: Sequence to Function through Folding, Binding, and Allostery. This comprehensive approach is poised to revolutionize how the scientific community and defense agencies understand and respond to biological threats, bolstering national readiness against both natural and synthetic biothreats.
