HyInv accelerates diverse protein sequence design from backbones

HyInv Technical Report

Machine Learning

Summary

Designing proteins means finding amino-acid sequences that fold well into a given shape. The team developed HyInv, a fast computer program that creates many diverse and well-fitting sequences for a protein backbone. It uses a neural network trained with a special method that combines guessing parts of the sequence and improving them based on 3D folding quality. HyInv is much faster than prior methods and works well on various protein design tasks, including enzyme and loop redesign.

What this means in practice

  • For protein engineers: Generate many diverse and well-folded protein sequences quickly for enzyme redesign or novel protein creation tasks.
  • For drug discovery teams: Produce libraries of protein variants with preserved catalytic sites or binding pockets to assist in drug target optimization.

Authors

Haowei Lin, Cai Zhou, Xiangxin Zhou

Abstract

Inverse folding designs amino-acid sequences for a given protein backbone, and screening needs many candidates that are both diverse and well folded. We present HyInv, a 1.96-million-parameter backbone-conditioned graph network with an encoder-decoder architecture, trained by Hyra 2.0, the research agent developed by our team: pretrained with random-order autoregressive sequence modeling, followed by structure-reward reinforcement learning. At inference, a grouped draft-and-rewrite sampler that uses separate draft and rewrite temperatures improves refolding TM-score at matched diversity. An optimized implementation with BF16 numerics designs 128 sequences for a 402-residue backbone in 50.6 ms on a single NVIDIA H20 GPU, about 52 times faster than the reference sampler, and an exact FP32 path that returns the reference sequences is 8.7 times faster. Across five benchmark datasets and seven dataset-predictor combinations, HyInv reaches a higher average scTM than all ten external baselines at comparable or higher sequence diversity, with large margins over ProteinMPNN and ProteinZero V2. The same default setting supports de novo backbones, pocket-constrained libraries, enzyme redesign with the catalytic site kept, and loop redesign. We have released our code and model at https://github.com/Tencent-Hunyuan/Hyra-results/tree/main/AI4Science/hyinv.