Abstract:Deep generative models (DGMs) have shown great success in the understanding and data-driven design of proteins. Variational autoencoders (VAEs) are a popular DGM approach that can learn the correlated patterns of amino acid mutations within a multiple sequence alignment (MSA) of protein sequences and distill this information into a low-dimensional latent space to expose phylogenetic and functional relationships and guide generative protein design. Autoregressive (AR) models are another popular DGM approach that typically lacks a low-dimensional latent embedding but does not require training sequences to be aligned into an MSA and enable the design of variable length proteins. In this work, we propose ProtWave-VAE as a novel and lightweight DGM, employing an information maximizing VAE with a dilated convolution encoder and an autoregressive WaveNet decoder. This architecture blends the strengths of the VAE and AR paradigms in enabling training over unaligned sequence data and the conditional generative design of variable length sequences from an interpretable, low-dimensional learned latent space. We evaluated the model's ability to infer patterns and design rules within alignment-free homologous protein family sequences and to design novel synthetic proteins in four diverse protein families. We show that our model can infer meaningful functional and phylogenetic embeddings within latent spaces and make highly accurate predictions within semisupervised downstream fitness prediction tasks. In an application to the C-terminal SH3 domain in the Sho1 transmembrane osmosensing receptor in baker's yeast, we subject ProtWave-VAE-designed sequences to experimental gene synthesis and select-seq assays for the osmosensing function to show that the model enables synthetic protein design, conditional C-terminus diversification, and engineering of the osmosensing function into SH3 paralogues.

Towards AI-designed genomes using a variational autoencoder

Designing minimal E. coli genomes using variational autoencoders

Variational autoencoder for design of synthetic viral vector serotypes

Can AI design life?

Accurate and General DNA Representations Emerge from Genome Foundation Models at Scale

ProtWave-VAE: Integrating Autoregressive Sampling with Latent-Based Inference for Data-Driven Protein Design

Deep-learning-based design of synthetic orthologs of SH3 signaling domains

Genetic Architect: Discovering Genomic Structure with Learned Neural Architectures

Learning to encode cellular responses to systematic perturbations with deep generative models

The design and engineering of synthetic genomes

Learning from and improving upon high-throughput screens for protein fitness with Generative AI - Application to BBB-crossing AAV design

Variational auto-encoding of protein sequences

Gene-language models are whole genome representation learners

The Genomic Code: The genome instantiates a generative model of the organism

Introducing Variational Autoencoders to High School Students

Inferring gene regulatory networks by hypergraph variational autoencoder

Sequence modeling and design from molecular to genome scale with Evo

VEGA is an interpretable generative model for inferring biological network activity in single-cell transcriptomics

Transformer Graph Variational Autoencoder for Generative Molecular Design

Meta-Learning an Evolvable Developmental Encoding

Generative artificial intelligence for de novo protein design