What ProteinMPNN is
ProteinMPNN is described in the IPD listing as a deep learning model for predicting amino acid sequences for protein backbones. In practice this means it addresses the sequence design step of protein engineering: given a backbone structure, the model proposes the amino acids that could fold into it. It appears in the catalogue alongside rfdiffusion, a generative model of protein backbones for protein binder design, which is also published by IPD.
- Publisher: IPD
- Task: amino acid sequence prediction for protein backbones
- Listed as a Model in the IPD collection
- Companion listing: rfdiffusion
Catalogue listing and access
ProteinMPNN is presented on the NVIDIA build site, which hosts models that can be tried through NVIDIA NIM APIs. The listing marks the model as downloadable, so it can be obtained in addition to being called as a hosted API endpoint. The IPD page groups it with one other model under the same publisher.
- Available via NVIDIA NIM APIs
- Marked Downloadable
- Part of the IPD publisher page with two models
Tags and usage signals
The catalogue applies a set of domain tags to ProteinMPNN that place it within biology and computational drug development workflows. The listing also reports usage and maintenance indicators, showing 6K API calls in the last 30 days and a last updated date in 2025.
- Tags: biology, nim, BioNemo, Protein Generation, Drug Discovery
- 6K API calls in the last 30 days
- Last updated in 2025
Sources
Last verified 02 Oct 2026. This entry is compiled from the public web pages listed above. Nothing here is stated that those pages do not, and each of them was read on the date shown.