Generative modeling for protein structures

By Namrata Anand, Po-Ssu Huang. 2018

176 citations recorded by the source corpus.

Subfields: Generative Models, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Generative Models

In Computational Biology & Bioinformatics

Citations in the corpus

Protein Structure and Sequence Generation with Equivariant Denoising Diffusion Probabilistic Models (2022, background): “Interpreting the pairwise distance matrix of a protein’s atoms as an image to be modeled with existing models has seen limited success because inconsistencies in the predictions lead to nontrivial errors when optimization routines are used to recover the final 3D structures [4].”
ProGen: Language Modeling for Protein Generation (2020, background): “This performance improves in settings with larger amino acid contexts and when ProGen is provided a larger number of conditioning tags, which highlights its potential for applications in providing viable, starting sequences for directed evolution or de novo protein design (Huang et al., 2016).”
Learning inverse folding from millions of predicted structures (2022, background): “…sequences for structures (Ingraham et al., 2019; Strokach et al., 2020; Anand-Achim et al., 2021; Jing et al., 2021b), generate backbone structures (Anand & Huang, 2018; Eguchi et al., 2020), jointly generate structures and sequences (An-ishchenko et al., 2021; Wang et al., 2021), or model…”
Protein sequence design with a learned potential (2022, methodology): “…it is now possible to build models that learn complex functions of protein sequence and structure, including models for protein backbone generation 18 – 20 and protein structure prediction 21,22 ; as a result, we were curious as to whether an entirely learned method could be used to design…”
Diffusion probabilistic modeling of protein backbones in 3D for the motif-scaffolding problem (2022, critical): “[1, 25, 39] propose generative adversarial networks, variational autoencoders, and energy-based models, respectively, on distance matrices, but these approaches (1) do not generate backbones compatible with a specified motif and (2) rely on an unwieldy optimization step to translate the distance mat”
Energy-based models for atomic-resolution protein conformations (2020, methodology): “Generative models of protein structures have also been proposed for generating protein backbones (Anand & Huang, 2018) and for the inverse protein folding problem (Ingraham et al., 2019).”
AlphaDesign: A de novo protein design framework based on AlphaFold (2021, background): “On the other side of the spectrum of neural network-based protein design, generative models have been explored for protein structure generation [60–62].”
Generative Models for Graph-Based Protein Design (2019, background): “In a complementary line of research, several deep and differentiable parameterizations of protein structure [26–29] have been recently proposed that could be used to generate, optimize, or validate 3D structures for input to sequence design.”
Deep Learning in Protein Structural Modeling and Design (2020, background): “reasonable Ramachandran plot Anand et al.(24)”
Protein sequence design with a learned potential (2020, background): “With the emergence of deep learning systems, which can learn patterns from high-dimensional data, it is now possible to build models that learn complex functions of protein sequence and structure, for example models for protein backbone generation [59] and protein structure prediction [60].”
Function-guided protein design by deep manifold sampling (2021, methodology): “…have focused on generative modeling for producing realistic protein structures — for example, using Generative Adversarial Networks (GAN) for creating pairwise distance maps [12] and variational autoencoders (VAE) for 3D coordinate generation of protein backbones [13] — and designing new sequences.”
Broadly applicable and accurate protein design by integrating structure prediction networks and diffusion generative models (2022, unknown): “…protein monomers is a longstanding challenge in protein design which is difficult to address with physically-based protein design methods due to the magnitude of the conformational sampling problem, and has been a primary test of deep learning based protein design approaches 2,3,5,10,12,24 .”