The Walter Lab at Harvard Medical School needed to map protein-protein interactions at proteome scale to study DNA repair and disease mechanisms, but manual structural biology work and false-positive model predictions made the project too slow and too large for traditional computing resources.
Reported impact
The Walter Lab produced nearly 1.7 million protein interaction predictions in three months, generated 40,000 high-confidence interactions among 300 genome-maintenance proteins, and created a practical path to broader structural interactome research that would otherwise have taken years or longer.
Harvard Medical School's Walter Lab built a multi-step AI pipeline for protein interaction discovery and scaled it on NVIDIA DGX Cloud. The workflow first uses a machine-learning tool called KIRC to identify protein pairs that merit deeper analysis, then applies a second model, the Structure Predictions and Omics-Informed Classifier (SPOC), to separate true and false protein-protein interaction predictions. To execute the structure-prediction stage at massive scale, the team ran ColabFold on the full NVIDIA DGX Cloud stack, using 32-node DGX clusters with 256 A100 GPUs obtained through the NAIRR pilot. This dedicated cloud environment gave the lab enough configurable GPU capacity to move from tens of thousands of predictions to 1.7 million PPI runs in three months, turning an otherwise years-long discovery process into a tractable research workflow.
KIRC first narrows the candidate protein pairs, SPOC filters the resulting predictions, ColabFold runs large-scale structure prediction on DGX Cloud A100 clusters, and the resulting interaction map feeds the lab's structural interactome and downstream disease-research hypotheses.
The Walter Lab conquered these limitations by developing a multi-step pipeline.
The researchers created a machine learning tool named KIRC, which identifies protein pairs that are worth in-depth analysis by AlphaFold.
The team adopted NVIDIA Accelerated Computing infrastructure and the full NVIDIA DGX Cloud stack—consisting of 32-node DGX clusters with 256 A100 GPUs.
The team then designed a second machine learning tool known as the Structure Predictions and Omics-Informed Classifier (SPOC), which is capable of separating true and false predictions of PPIs.
This NVIDIA solution allowed researchers to run 1.7 million PPI predictions on ColabFold in the span of three months.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Oct 6, 2026 University of Pennsylvania Penn needed shared AI infrastructure that could support researchers across many disciplines without weeks-long wait times, fragmented storage, or the cost barriers that kept frontier-scale training and inference off… 1.5 years of compute time reduced to three days for a 40,000-document project NVIDIA DGX B200 systems · NVIDIA DGX SuperPOD · NVIDIA InfiniBand NVIDIA Sample Aug 27, 2026 Texas A&M University Texas A&M needed to give researchers across seven institutions access to shared GPU infrastructure for AI and simulation, but isolated workstations, slow grant-funded clusters, and long queues were limiting… $1M saved on a single project DDN high-performance storage · NVIDIA Base Command Manager · NVIDIA DGX SuperPOD NVIDIA Sample Aug 3, 2026 StudyFetch StudyFetch needed to keep AI tutoring affordable while scaling live lecture transcription and personalized learning for millions of students, but off-the-shelf cloud transcription created cost and latency ceilings that… 100M+ learning interactions NVIDIA H100 GPUs · NVIDIA L40S GPUs · NVIDIA NIM 