variational autoencoder

Using deep generative models for simultaneous representational and predictive modeling of brain and behavior: A graded supervised-to-unsupervised modeling framework.

This paper develops a framework combining generative and discriminative models to evaluate brain-behavior mappings, showing how it captures diverse relationships and identifies assumption violations.

The use of variational autoencoders to characterise the heterogeneous subpopulations that arise due to antibiotic treatment

This study uses a Bayesian Variational Autoencoder to analyze bacterial single-cell RNA sequencing data and identify subpopulations responding to antibiotic treatment.

Multimodal normative modeling in Alzheimer Disease with introspective variational autoencoders

This paper introduces a multimodal introspective VAE for Alzheimer's Disease normative modeling, improving reconstruction of healthy controls and outlier detection in neuroimaging data.

A generative deep learning method for global species distribution prediction

This paper introduces EcoVAE, a generative deep learning model for predicting global species distributions, achieving high precision and outperforming traditional methods.

inVAE: Conditionally invariant representation learning for generating multivariate single-cell reference maps

inVAE improves cell state identification by disentangling biological and technical variation in single-cell data, outperforming existing methods across multiple human tissues.

Predicting proprioceptive cortical anatomy and neural coding with topographic autoencoders

The paper develops a topographic autoencoder to model proprioceptive cortical coding and predicts neural manifold structures from natural movement data.

Balancing Locality and Reconstruction in Protein Structure Tokenizer

This paper introduces a VQ-VAE-based protein structure tokenizer that balances reconstruction and retrieval, improving alignment with sequence models for better structure prediction.

Towards AI-designed genomes using a variational autoencoder

This paper introduces a variational autoencoder to model bacterial genomes, demonstrating its ability to reconstruct and generate functional genome vectors with high accuracy.

Synthetic augmentation of cancer cell line multi-omic datasets using unsupervised deep learning

The paper introduces a deep learning model to augment cancer cell line multi-omic data, increasing the number of profiles and uncovering drug resistance mechanisms.

Identifying mechanisms and therapeutic targets in muscle using Bayesian parameter estimation with conditional variational autoencoders

This paper uses a Conditional Variational Autoencoder to predict muscle twitch parameters from simulated and experimental data, identifying therapeutic targets for a cardiac mutation.

Generative Models for Prediction of Non-B DNA Structures

The paper evaluates generative models for improving non-B DNA structure prediction through data augmentation, finding diffusion models perform best for Z-DNA and H-DNA while WGAN excels for G-quadruplexes.

SIMPL: Scalable and hassle-free optimisation of neural representations from behaviour

SIMPL optimizes neural representations of behaviour by iteratively refining latent variables and tuning curves, revealing more precise spatial encoding than traditional methods.

SMAS: Structural MRI-Based AD Score using Bayesian VAE

This paper introduces SMAS, a Bayesian VAE-based AD score from structural MRI data, showing strong correlations with cognitive decline and brain atrophy markers, and outperforming existing measures in predicting AD.

Discovering uncertainty: Bayesian constitutive artificial neural networks

This paper introduces Bayesian constitutive neural networks to model soft matter systems, showing they can discover interpretable models, quantify uncertainties, and distinguish between healthy and diseased arteries.

Prediction of context-specific regulatory programs and pathways using interpretable deep learning

The paper introduces COBRA, an interpretable deep learning model for scRNA-seq data that disentangles covariate effects, and demonstrates its application on interferon-stimulated mouse datasets.

Dynamic Causal Modeling in Probabilistic Programming Languages

This paper develops Bayesian inference methods for Dynamic Causal Modelling using probabilistic programming languages to improve neural dynamics prediction.

Disentangling cellular heterogeneity into interpretable biological factors through structured latent representations

DRVI is a deep generative model that learns interpretable biological factors from single-cell omics data without supervision, achieving both disentanglement and integration performance.

Unified Analysis of Multi-order Tensors for Integrative Molecular Profiling

The paper introduces MUSIC, a Bayesian tensor decomposition framework for integrating multi-modal molecular data, showing it reveals interpretable patterns across patients, cell types, and modalities.

A Variational Graph Partitioning Approach to Modeling Protein Liquid-liquid Phase Separation

This paper introduces a Graph Partitioned GNN (GP-GNN) to predict protein liquid-liquid phase separation by focusing on subgraphs containing intrinsically disordered regions.

VAPOR: A variational autoencoder with transport operators to disentangle cellular gene expression dynamics of co-occurring biological processes in time and space

VAPOR disentangles co-occurring biological processes in gene expression dynamics using a variational autoencoder with transport operators, applied to single-cell and spatial transcriptomics data.

CopyMix: Mixture Model Based Single-Cell Clustering and Copy Number Profiling using Variational Inference

CopyMix jointly infers cell clusters and copy number profiles from single-cell DNA sequencing data, avoiding clustering artifacts and parameter tuning.

Designing minimal E. coli genomes using variational autoencoders

This study uses variational autoencoders to generate minimal E. coli genomes, finding 6 out of 100 sampled genomes were viable in a computational model.

The 3D point spread function can be autoencoded in latent space

This paper shows that the 3D point spread function of a fluorescent emitter can be directly accessed in the latent space of an autoencoder trained on sample images without z-reference data.

Sparse-Coding Variational Auto-Encoders

This paper introduces a sparse-coding variational auto-encoder (SVAE) that improves upon sparse coding by using a probabilistic model with a deep neural network, achieving better performance on natural image data.

Exploring "dark matter" protein folds using deep learning

This paper introduces a deep learning framework called Genesis to learn protein structural patterns and demonstrates its ability to design novel protein folds with high protease resistance.

BEATRICE: Bayesian Fine-mapping from Summary Datausing Deep Variational Inference

BEATRICE identifies causal variants from GWAS summary data using a Bayesian framework, outperforming baselines in coverage and finding the APOEε2 allele in Alzheimer's disease.

Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping

The paper introduces four VAE-based deconfounding frameworks for multi-omics clustering, showing they reduce confounding effects and improve biological clustering.

DeepResBat: deep residual batch harmonization accounting for covariate distribution differences

DeepResBat improves MRI data harmonization by addressing covariate distribution differences, outperforming existing methods while avoiding false associations.

Joint Trajectory Inference for Single-cell Genomics Using Deep Learning with a Mixture Prior

This paper introduces VITAE, a deep learning method for trajectory inference in single-cell genomics that improves accuracy and uncertainty quantification across datasets.

LineageVAE: Reconstructing Historical Cell States and Transcriptomes toward Unobserved Progenitors

LineageVAE reconstructs historical cell states and transcriptomes from scRNA-seq data to infer unobserved progenitor cell states.