Model Card for Pinal: Programmable design of functional proteins from natural language
Pinal is an advanced protein design framework that translates human design intent into novel protein sequences. It utilizes a two-stage process: first generating protein structures from language instructions, then designing sequences based on those structures. With a substantial parameter count of 16 billion and trained on a diverse dataset comprising 1.7 billion protein-text pairs, Pinal demonstrates marked improvements in both performance and generalization capabilities for novel protein structures.
For more information, please refer to our paper, Programmable design of functional proteins from natural language, published in Nature (2026).
SaProt-T(ext) and SaProt-O(mni)
SaProt-T was specifically developed for protein reverse folding within the Pinal framework, whereas SaProt-O is trained with configurable inputs (optional textual prompts and structural features) to accommodate real-world research scenarios.
Usage
Please refer to the README.md in the Github repository for details on how to use the model.
- Downloads last month
- -