IBS-privacy

IBS-privacy evaluates vulnerabilities in genetic privacy arising from identical-by-state (IBS) matching between uploaded genomes and genealogical databases to quantify risks of genotype recovery and imputation-based attacks.

Key Features:

• Risk Assessment of IBS Matches: Analyzes how IBS matching methods used by DTC services can expose users' genotypes and highlights vulnerabilities that arise when unphased genotypes are employed.
• Attack Simulation and Demonstration: Provides a proof-of-concept demonstration that uploading approximately 900 publicly available genomes could allow an adversary to recover significant portions of genomes in a database.
• Genome-wide Imputation Analysis: Assesses the risk of genome-wide imputation attacks, showing that around 100 falsified datasets can be used to infer detailed genetic information about all users in databases that use unphased genotypes for IBS detection.
• Database Vulnerability Comparison: Compares different DTC genealogical services based on their methods for identifying and reporting IBS segments to indicate varying levels of vulnerability.
• Mitigation Strategies: Presents practical recommendations to reduce the exploitation of identified vulnerabilities in genealogical databases.

Scientific Applications:

• Genetic privacy risk assessment: Supports researchers and developers evaluating how uploads to DTC genetic genealogy services can compromise genotype privacy and informing the development of more secure IBS detection and reporting methods.

Methodology:

Evaluating existing datasets from DTC services to identify common patterns of IBS matches. Simulating attacks using techniques like IBS tiling to demonstrate how adversaries might reconstruct genotypes. Comparing genealogical databases by analyzing their IBS detection and reporting methods to determine susceptibility to privacy breaches.