K-Fold

K-Fold is a biomolecular foundation model that predicts binding-induced conformational changes. Through an apo-to-holo diffusion bridge, it models how unbound molecules assemble and change shape upon binding, including GPCR and kinase systems.

Preprint will be available soon.

Model parameters

K-Fold uses pretrained AtlasLM for protein sequence representations and TriProRep for protein structure representations. The parameters for these models and K-Fold are downloaded automatically on first use from Hugging Face.

Installation

K-Fold requires Python 3.11 or later.

Install from PyPI:

pip install kfold

Or install from source:

git clone https://github.com/kaist-ai-bio/kfold.git
cd kfold
pip install -e .

Inference

Run predictions from a YAML or JSON query file, or a directory of query files for bulk runs:

kfold --input examples/8and.yaml --out-dir predictions/ --seeds 42

See kfold --help, the inference guide, or the Python API guide for details.

By default, K-Fold prepares apo structures with AtlasFold, then runs complex structure prediction. You can also provide apo structures from experiments or other prediction tools (e.g., AlphaFold2).

Efficient inference with multiple seeds: For multi-seed inference, use --share-apo-seeds to generate an apo ensemble for each protein entry and reuse it across all inference seeds. This is particularly useful for relatively rigid apo structures or runs with many inference seeds; see automatic apo generation.

kfold --input examples/ --out-dir predictions/ --share-apo-seeds 1 2 3 --seeds 1 2 3 4 5 6 7 8 9 10

Multi-stage inference: Use --stage apo to prepare apo structures only, or --stage complex to predict complexes from prepared apos:

kfold --stage apo --input examples/ --out-dir predictions/ --seeds 42
kfold --stage complex --input examples/ --out-dir predictions/ --seeds 42

Kernel selection: Use --kernel {auto,triton,cuequiv,torch} to select the backend for AtlasFold apo preparation and K-Fold complex prediction. The default, auto, prefers Triton, then cuEquivariance, then PyTorch, depending on availability.

Training

See the training guide for data preparation, training commands, and configuration.

Citation

TBA

Acknowledgements

K-Fold was developed at KAIST as part of the K-Fold initiative supported by the Ministry of Science and ICT (MSIT), Republic of Korea.

Members of Team KAIST are listed below (alphabetical order):

• Project management: Hyeongwoo Kim^3,†
• Engineering lead: Seonghwan Seo^3,†
• K-Fold architecture: Seokhyun Moon^3,†, Jun Hyeong Kim³, Shinwoo Kim³, Minha Park³, Jisu Seo³, Mingyeong Shin³, Wonho Zhung³
• Protein structure encoder: Hyosoon Jang^1,†, Taewon Kim^1,†, Hyunjin Seo¹
• RNA sequence encoder: Dongki Kim^1,†, Jun Hyeong Kim^1,†, Jinheon Baek¹, Jaehyeong Jo¹
• Training data preparation: Yeongnam Bae^2,†, Woosung Jeon^2,†, Joongwon Lee^3,†, Junyup Lee^2,†, Yunsu Shin^2,†, Eugene Choi², Jeong Hun Choi², Hyeongyu Han², Calvin Samuel²
• Kernel optimization: Youngchan Kim⁴
• Supervision: Sungsoo Ahn¹, Dongsu Han^4,1, Sung Ju Hwang¹, Ho Min Kim², Woo Youn Kim³, Gyu Rie Lee², Byung-Ha Oh²

^† Core contributor; ¹ KAIST AI; ² KAIST Biological Sciences; ³ KAIST Chemistry; ⁴ KAIST Electrical Engineering.

We thank our collaborators at HITS for their contributions to K-Fold.

License

Copyright © 2026 Korea Advanced Institute of Science and Technology (KAIST).

K-Fold source code and model weights are licensed under the Apache License 2.0.