SE(3)-Equivariant Energy-based Models for End-to-End Protein Folding
By Jiaxiang Wu, Tao Shen, Haidong Lan, Yatao Bian, Junzhou Huang. 2021
11 citations recorded by the source corpus.
Accurate prediction of protein structures is critical for understanding the biological function of proteins. Nevertheless, most structure optimization methods are built upon pre-defined statistical energy functions, which may be sub-optimal in formulating the conformation space. In this paper, we propose an end-to-end approach for protein structure optimization, powered by SE(3)-equivariant energy-based models. The conformation space is characterized by a SE(3)-equivariant graph neural network, with substantial modifications to embed the protein-specific domain knowledge. Furthermore, we introduce continuously-annealed Langevin dynamics as a novel sampling algorithm, and demonstrate that such process converges to native protein structures with theoretical guarantees. Extensive experiments indicate that SE(3)-Fold achieves comparable structure optimization accuracy, compared against state-of-the-art baselines, with over 1-2 orders of magnitude speed-up.
Subfields: Computational Theory and Mathematics, Molecular Biology
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Theory and Mathematics
- Tommi S. Jaakkola
- Regina Barzilay
- Octavian-Eugen Ganea
- Shigenori Tanaka
- Shuya Nakata
- Yoshiharu Mori
- Andreas Krause
- Charlotte Bunne
In Molecular Biology
- Tommi S. Jaakkola
- Regina Barzilay
- Shuya Nakata
- Shigenori Tanaka
- Yoshiharu Mori
- Andreas Krause
- Charlotte Bunne
- Octavian-Eugen Ganea
Citations in the corpus
- End-to-end protein–ligand complex structure generation with diffusion-based generative models (2022, background): “In particular, diffusion-based generative models [23–26] have demonstrated a capacity for high-quality synthesis in many domains, including conformation generation of small molecules [27–30] and proteins [31–33].”
- Crystal Diffusion Variational Autoencoder for Periodic Material Generation (2021, methodology): “In addition, score-based and energy-based models have also been used for molecular graph generation (Liu et al., 2021) and protein folding (Wu et al., 2021).”
- Independent SE(3)-Equivariant Models for End-to-End Rigid Protein Docking (2021, background): “for molecular structures (Fuchs et al., 2020; Hutchinson et al., 2020; Wu et al., 2021; Jumper et al., 2021; Ganea et al., 2021) and different types of 3D point clouds (Thomas et al.”