Privacy-Preserving Technologies in Data

Evaluation of synthetic data generation for longitudinal 6-year follow-up subclinical cardiovascular disease data using GAN models

This study evaluates CTGAN and TGAN for generating synthetic longitudinal CVD data, finding CTGAN offers better fidelity than TGAN but has lower performance in temporal fidelity and for younger age groups.

Evaluating in-context learning with prompting regimes for transformer-based synthetic health tabular data generation

This study evaluates zero shot, one shot, and few shot prompting regimes for transformer-based synthetic health tabular data generation using small models, finding zero shot improves fidelity but reduces privacy.

FedEdgeR: federated and privacy-preserving edgeR for differential gene expression analysis

FedEdgeR enables federated differential gene expression analysis while preserving privacy, outperforming meta-analysis methods and matching pooled edgeR results.

Unlocking Sensitive Data with SPHERE in the Age of AI

SPHERE enables the use of sensitive datasets in AI and open science by creating synthetic twins that preserve statistical properties and privacy.

NeuroFLAME: A Scalable, Privacy-Preserving Federated Framework for Secure, Reproducible, and Multi-Site Neuroimaging Analysis

NeuroFLAME provides a secure, scalable federated platform for multi-site neuroimaging analysis, achieving consistency with centralized methods while preserving data privacy and enabling institutional compliance.

The Synthetic Fidelity-Stability Framework (SFSF): A Systematic Multi-Dimensional Benchmark of Synthetic Clinical Laboratory Data Generators

The paper benchmarks 17 synthetic data generators for clinical lab data, finding no single model excels across all metrics, and highlights ForestDiffusion's best privacy-utility balance.

Membership Inference on Synthetic Single-Cell Genomic Data

This paper demonstrates that synthetic scRNA-seq data generated by leading methods can be attacked to infer donor membership, showing privacy risks and evaluating noise-based defenses.

Synthetic RNA-seq cohorts for data sharing: a discovery-aware benchmark at transcriptome scale

The paper benchmarks synthetic RNA-seq generation methods across real-world cohorts to evaluate biological fidelity and privacy risks.

Distribution-Aware Federated Learning for Diabetes Prediction Using Tabular Clinical Data Under Non-IID and Class-Imbalanced Settings

This paper introduces DA-FL, a federated learning method that improves diabetes prediction in non-IID and class-imbalanced settings by incorporating class distribution awareness.

Towards Useful and Private Synthetic Omics: Community Benchmarking of Generative Models for Transcriptomics Data

The paper benchmarks generative models for synthetic bulk RNA-seq data, finding trade-offs between utility, biological preservation, and privacy risks across different methods.

BioVault: A privacy-first data visitation platform for equitable global collaboration in biomedicine

BioVault enables privacy-first biomedical collaboration by allowing analysis code to visit data locally, avoiding data transfer and preserving privacy and data sovereignty.

Genomic privacy risks in GWAS summary statistics

This study shows that combining GWAS summary statistics with phenotype data can leak genomic information, with higher risk for non-European populations.

FMed-Diffusion Federated Learning on Medical Image Diffusion

This paper proposes FMedDiffusion, a federated learning approach for medical image diffusion models, achieving performance comparable to traditional training on three medical image datasets.

Validation of ten federated learning strategies for multi-contrast image-to-image MRI data synthesis from heterogeneous sources

The paper validates ten federated learning strategies for synthesizing multi-contrast MRI data from heterogeneous sources, finding that FedBAdam outperforms other methods in accuracy and stability.

NoisyFlow: Differentially Private Optimal Transport Using Neural Networks for Secure Biomedical Data Sharing

NoisyFlow enables secure biomedical data sharing by aligning data distributions across institutions using differentially private optimal transport with neural networks.

Bridging Earth and Space: A Flexible and Resilient Federated Learning Framework Deployed on the International Space Station

This paper presents a federated learning framework deployed on the International Space Station to train models using data from Earth and the ISS without transferring raw data.

FedPyDESeq2: a federated framework for bulk RNA-seq differential expression analysis

FedPyDESeq2 enables privacy-preserving differential expression analysis on siloed bulk RNA-seq data, achieving results comparable to pooled data analysis and outperforming meta-analysis approaches.

Safeguarding Privacy in Genome Research: A Comprehensive Framework for Authors

This paper introduces a framework to evaluate privacy risks in genomic data sharing by assessing compliance with NIH policies and identifying vulnerabilities through self-reported practices.

A cautionary tale on the cost-effectiveness of collaborative AI in real-world medical applications.

This study benchmarks the accuracy and cost-effectiveness of federated learning and consensus-based learning in medical data analysis, finding that consensus-based learning is more cost-effective with comparable accuracy.

Privacy-preserving impact evaluation using Difference-in-Differences

The paper develops a federated Difference-in-Differences method for privacy-preserving causal impact evaluation using the DataSHIELD platform, demonstrating its effectiveness on simulated and real-world data.

ProxyTyper: Generation of Proxy Panels for Privacy-aware Outsourcing of Genotype Imputation

ProxyTyper generates proxy panels to protect genetic data privacy during genotype imputation with minimal accuracy loss.

dsSurvival 2.0: Privacy enhancing survival curves for survival models in the federated DataSHIELD analysis system

The paper introduces an enhanced dsSurvival package for generating privacy-preserving survival curves within the DataSHIELD federated analysis system.

Fed-ComBat: A Generalized Federated Framework for Batch Effect Harmonization in Collaborative Studies

Fed-ComBat harmonizes batch effects in decentralized data, showing comparable results to centralized methods for linear and nonlinear cases.

FedDP: Secure Federated Learning for Disease Prediction with Imbalanced Genetic Data

FedDP improves disease prediction from imbalanced genetic data by combining federated learning with differential privacy and a weighted minmax loss.

Secure and federated linear mixed model association tests

This paper introduces SafeGENIE, a secure and efficient algorithm for LMM-based GWAS that enables collaborative analysis without data sharing.

Efficient Differentially Private Methods for a Transmission Disequilibrium Test in Genome Wide Association Studies

The paper introduces efficient differentially private methods for a transmission disequilibrium test in genome-wide association studies, showing significant speed improvements over existing methods.

Privacy-preserving quality control of neuroimaging datasets in federated environments

The paper introduces decentralized quality control tools for neuroimaging data in federated environments, showing promising results in visualization and outlier detection while preserving privacy.

CanDIG: Secure Federated Genomic Queries and Analyses Across Jurisdictions

CanDIG enables secure, federated genomic and health data analysis across jurisdictions while maintaining data locality and privacy.

dsSwissKnife: An R package for federated data analysis

dsSwissKnife enables federated data analysis in R, supporting biomarker discovery across harmonized datasets from multiple observational cohorts.

Privacy preserving storage of sequenced genomic data

The paper introduces Varlock, a method to securely store genomic data by masking personal alleles using population frequencies, ensuring privacy while preserving useful data for research.