variational autoencoder

Unified Network Spectroscopy: Inferring Reduced Biological Dynamics from Fluctuations under Structured Noise and Broken Ergodicity

This paper introduces Unified Network Spectroscopy (UNS), a method to infer biological dynamics from fluctuations, addressing issues with noise structure and ergodicity.

LGTM: Gaussian Process Modulated Neural Topic Modeling for Longitudinal Microbiome

LGTM couples a neural topic model with additive basis-function Gaussian processes, learning microbial co-abundance topics and how covariates relate to their trajectories in longitudinal gut microbiome data.

Setting the SCENE for Interpretable Cell-Gene Embeddings in Single-Cell RNA-seq

This paper introduces SCENE, a probabilistic model for embedding cells and genes in scRNA-seq data, achieving state-of-the-art performance and interpretable visualization.

Modelling interpretable patient-level representationsfrom structured and simple multimodal data

The paper introduces FACTMx, a variational framework that jointly models structured and simple patient-level data to learn interpretable representations, showing improved performance in simulations and applications to cancer and cardiovascular cohorts.

MOFTy: Multimodal Gaussian Process Factor Analysis with Numerical Information Field Theory

MOFTy disentangles spatial structures and recovers cross-modal patterns in multi-omics data by separating smooth gradients from fine-scale heterogeneity.

Learning Discrete Cell and Niche Codes from Spatial Transcriptomics Using Dual Residual Vector Quantization

SQUINT learns discrete cell and niche codes from spatial transcriptomics data using residual vector quantization, outperforming baselines in identification and cross-section integration.

Learning with Recurrence Geometric AI in Spatial Transcriptomics

This paper introduces RGAI, a geometric deep-learning framework for cross-domain analysis of spatial transcriptomics, enabling alignment of latent tissue states across diverse datasets.

OMIDIENT: Multiomics Integration for Cancer by Dirichlet Auto-Encoder Networks

OMIDIENT is a multi-omics integration method that uses Dirichlet auto-encoders to improve clustering, classification, and data reconstruction across cancer types.

CoLa-VAE: A Cell-Cell Communication-Aware Variational Autoencoder for Representation Learning and Expression Denoising

CoLa-VAE improves denoising and representation learning in single-cell data by integrating cell-cell communication information, enhancing downstream analyses like differential communication detection and spatial deconvolution.

ARCADIA Reveals Spatially Dependent Transcriptional Programs through Integration of scRNA-seq and Spatial Proteomics

ARCADIA integrates scRNA-seq and spatial proteomics to reveal spatially dependent transcriptional programs without cell barcode pairing or feature-level correspondence.

Deep Learning-Based Reconstruction: Model Comparison for Variable-Density GRAPPA 1H MRSI

This paper compares deep learning models for improving 1H MRSI reconstruction, finding that U-Net outperforms MultiNet in reducing lipid artifacts and improving metabolite map fidelity.

Latent Representations of Early Brain Development: A Multivariate Normative Model of Brain Structure and Behaviour

This paper uses a semi-supervised autoencoder to create multivariate normative models of brain structure and behaviour in children, showing associations between structural deviations and cognitive/behavioural differences.

The hippocampus as a small-world cognitive map

The paper uses a geometry-aware autoencoder to analyze hippocampal activity in mice, revealing a small-world network structure that enables both local fidelity and global searchability in cognitive maps.

bMAE: Masked Autoencoder Latent Representations for Bulk RNA-seq Tissues

The paper introduces a masked autoencoder for bulk RNA-seq data to learn latent representations that outperform existing methods in tissue discrimination and structure preservation.

petVAE: A Data-Driven Model for Identifying Amyloid PET Subgroups Across the Alzheimer's Disease Continuum

petVAE identifies biologically meaningful subgroups along the Alzheimer's disease continuum using a deep learning model trained on amyloid PET scans.

Deep Computational Anatomy via Latent-Aligned Multiview Normalizing Flows

The paper introduces LAMNr flows for multimodal data analysis, enabling shared latent subspaces and cross-view imputation, with applications in biological imaging.

BundleWarp: Enhancing White Matter Tractometry and Morphometry with Precise Neuronal Mapping using Streamline-based Nonlinear Registration

BundleWarp improves white matter tractometry by aligning tracts while preserving anatomical features, enhancing sensitivity to disease-related changes in conditions like Alzheimer's and Parkinson's.

Decoding Anadara shell morphology with deep learning

This paper introduces Morpho-VAE, a deep learning framework for analyzing shell morphology, which outperforms traditional methods in species classification and reveals morphological patterns in Anadara shells.

Revisiting Reconstruction Likelihood: Variational Autoencoders for Biological and Biomedical Data Clustering

This paper shows that variational autoencoders can effectively cluster data by encoding cluster assignments in their latent space, using reconstruction likelihood to identify typical samples and anomalies.

Integrating single-cell and single-nucleus datasets improves bulk RNA-seq deconvolution

This study improves bulk RNA-seq deconvolution by integrating single-cell and single-nucleus datasets through filtering and transformation strategies, achieving near-scRNA-seq accuracy.

MultiGAI: Global Attention-Based Integration of Single-Cell Multi-Omics with Batch Effects Correction

MultiGAI corrects batch effects in single-cell multi-omics data while preserving biological signals, showing effectiveness in integration and correction tasks.

Partially Shared Multi-Modal Embedding Learns Holistic Representation of Cell State

APOLLO learns shared and modality-specific information from multi-modal single-cell data to improve understanding of cell state and regulatory mechanisms.

ABaCo: Addressing Heterogeneity Challenges in Metagenomic Data Integration with Adversarial Generative Models

ABaCo integrates metagenomic data from multiple studies, corrects technical heterogeneity, and outperforms existing methods while preserving biological signals.

Skip-Zeros Variational Inference in the Million-Cell Era of Single-Cell Transcriptomics

UNISON is a scalable framework for single-cell transcriptomics that uses skip-zeros variational inference to efficiently analyze large, sparse datasets while preserving interpretability.

Conditioned Graph Reconstruction of Brain Functional Network Connectivity Reveals Interpretable Latent Axes of Sex and Fluid Intelligence

The paper introduces a conditional graph variational autoencoder to model brain functional connectivity, reconstructing network patterns conditioned on sex and fluid intelligence while preserving interpretable latent features.

A hyperparameter benchmark of VAE-based methods for scRNA-seq batch integration

This study benchmarks VAE-based methods for scRNA-seq batch integration, finding dataset-dependent performance trade-offs and recommending hyperparameter settings for optimal results.

Latent-centric Isotropic Resolution Enhancement for Expansion Microscopy Imaging via Neural Compression and Self-supervised Learning

This paper introduces a self-supervised framework for isotropic super-resolution and compression in expansion microscopy, enabling efficient storage and analysis of whole-organ datasets.

Deep Variational Autoencoders for Population Genetics

This paper applies variational autoencoders to genomic data to capture population structure, enable dimensionality reduction, and improve data compression across diverse populations.

Learning the Shape of Evolutionary Landscapes: Geometric Deep Learning Reveals Hidden Structure in Phenotype-to-Fitness Maps

This paper introduces a geometric deep learning method to model phenotype-to-fitness maps, showing it outperforms linear methods in capturing nonlinear structure and predicting fitness landscapes.

Brain functional network connectivity interpolation characterizes the neuropsychiatric continuum and heterogeneity

This paper introduces a variational autoencoder framework to interpolate functional network connectivity data across neuropsychiatric disorders, revealing continuous gradients and heterogeneity.