Privacy-Preserving Technologies in Data

Evaluating anonymized genome re-identification using polygenic predictions and its implications for data privacy

This paper evaluates the feasibility of re-identifying individuals using polygenic predictions and finds that real-world effectiveness is limited due to low priors and poor sensitivity for sensitive genotypes.

Privacy-Preserving Pangenome Graphs

PanMixer reduces re-identification risks in pangenome graphs while preserving utility for downstream genomic analyses.

Privacy-Preserving Visualization of Brain Functional Connectivity

This paper explores privacy-preserving visualization techniques for brain functional connectivity using differential privacy to protect participant data while maintaining visual utility.

A simple demonstration of a privacy-preserving de-centralised genotype imputation workflow

The paper demonstrates a privacy-preserving decentralized genotype imputation workflow using a federated approach, showing it can protect sensitive data while enabling imputation on plain-text files.

Private Information Leakage from Polygenic Risk Scores

This paper shows that polygenic risk scores can be used to infer genotypes and de-anonymize individuals, and proposes a framework to assess and mitigate privacy risks when sharing PRS values.

PRISM-G: an interpretable privacy scoring method for assessing risk in synthetic human genome data

PRISM-G quantifies privacy risk in synthetic genome data by assessing genetic proximity, familial patterns, and trait-linked exposure, showing vulnerabilities vary across models and marker densities.

Federated Learning for ICU Mortality Prediction: Balancing Accuracy and Privacy in a Multi-Hospital Setting

This study develops a federated learning framework for ICU mortality prediction, showing improved accuracy and privacy compared to local models using multi-hospital data.

dGAMLSS: An exact, distributed algorithm to fit Generalized Additive Models for Location, Scale, and Shape for privacy-preserving population reference charts

The paper introduces dGAMLSS, a distributed algorithm for fitting GAMLSS models to construct population reference charts without sharing patient data.

A Reinforcement Learning-based Approach for Dynamic Privacy Protection in Genomic Data Sharing Beacons

This paper introduces a reinforcement learning-based method to dynamically protect genomic data sharing beacons against membership inference attacks by adapting strategies in real time.

FedNolowe: A Normalized Loss-Based Weighted Aggregation Strategy for Robust Federated Learning in Heterogeneous Environments

FedNolowe improves federated learning stability by weighting client contributions based on normalized losses, achieving state-of-the-art performance with reduced computational complexity.

Federated deep learning enables cancer subtyping by proteomics

ProCanFDL uses federated deep learning to subtype cancers using proteomic data while preserving privacy, achieving performance gains over local models.

Beacon Reconstruction Attack: Reconstruction of genomes in genomic data-sharing beacons using summary statistics

This paper identifies a privacy vulnerability in genomic data-sharing beacons, demonstrating genome reconstruction using summary statistics and SNP correlations, achieving a higher F1-score than baseline methods.

Secure and Federated Genome-Wide Association Studies for Biobank-Scale Datasets

This paper introduces SF-GWAS, a secure federated framework for GWAS that enables efficient and accurate analysis of private genomic data across institutions.

A comprehensive experimental comparison between federated and centralized learning

This paper compares federated and centralized learning through experiments, finding similar performance and federated learning's robustness to data imbalances and high dimensionality.

gLinDA: global differential abundance analysis ofmicrobiomes

gLinDA analyzes microbiome data using swarm learning to detect differential abundance without sharing raw patient data.

Efficient Federated Learning for distributed NeuroImaging Data

The paper proposes a decentralized sparse federated learning approach for neuroimaging data to reduce communication overhead while maintaining data privacy.

Secure Discovery of Genetic Relatives across Large-Scale and Distributed Genomic Datasets

SF-Relate identifies genetic relatives across distributed genomic datasets using a privacy-preserving federated approach with high accuracy and efficiency.

FedGMMAT: Federated Generalized Linear Mixed Model Association Tests

FedGMMAT enables privacy-preserving genetic association testing across institutions using federated learning and homomorphic encryption.

Federated Learning for Predicting Compound Mechanism of Action Based on Image-data from Cell Painting

This work uses federated learning to predict compound mechanism of action from cell painting image data, showing its effectiveness in handling data heterogeneity across institutions.

Federated Learning for multi-omics: a performance evaluation in Parkinson's disease

This study evaluates federated learning models for multi-omics Parkinson's disease prediction, finding they perform comparably to centrally trained models with minor performance differences.

Efficacy of federated learning on genomic data: a study on the UK Biobank and the 1000 Genomes Project

This study evaluates federated learning's effectiveness on genomic data using UK Biobank and 1000 Genomes Project datasets, showing performance close to centralized models despite data heterogeneity.

PPML-Omics: a Privacy-Preserving federated Machine Learning method protects patients' privacy in omic data

PPML-Omics is a privacy-preserving federated learning method that balances privacy protection and model utility across three omic data tasks.

Audit to Forget: A Unified Method to Revoke Patients' Private Data in Intelligent Healthcare

The paper introduces AFS, a tool to revoke patients' private data from pre-trained models by auditing usage and purifying knowledge, tested across multiple tasks and datasets.

Accessibility of covariance information creates vulnerability in Federated Learning frameworks

This study introduces an attack that reconstructs privacy-protected data in FL frameworks by exploiting covariance information and proposes defense strategies using differential privacy.

Assessing transcriptomic re-identification risks using discriminative sequence models

This paper introduces a discriminative sequence model to assess transcriptomic re-identification risks by improving linking attack accuracy across datasets.

Privacy-Preserving Statistical Analysis of Genomic Data using Compressive Mechanism with Haar Wavelet Transform

This paper introduces two privacy-preserving methods for genomic data analysis using compressive mechanisms and Haar wavelet transforms, achieving differential privacy with high utility.

dsMTL - a computational framework for privacy-preserving, distributed multi-task machine learning

The paper introduces dsMTL, a framework for privacy-preserving, distributed multi-task machine learning, and evaluates its performance in identifying biological signatures across distributed datasets.

dsSurvival: Privacy preserving survival models for federated individual patient meta-analysis in DataSHIELD

The paper introduces dsSurvival, a tool for privacy-preserving survival analysis in federated settings, enabling meta-analysis of Cox regression models without sharing individual data.

Truly Privacy-Preserving Federated Analytics for Precision Medicine with Multiparty Homomorphic Encryption

The paper introduces FAMHE, a privacy-preserving federated analytics system using multiparty homomorphic encryption, enabling accurate analysis of distributed medical data without sharing patient-level information.

Swarm Learning as a privacy-preserving machine learning approach for disease classification

This paper introduces Swarm Learning, a decentralized machine learning approach for disease classification using transcriptomic data, demonstrating its effectiveness in developing classifiers for COVID-19, tuberculosis, and leukemias while preserving privacy.