Differentially Private Linkage Analysis with TDT — the case of two affected children per family
Akito Yamamoto, Tetsuo Shibuya · 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · 2021
Statistical analyses of datasets containing genomic information is essential for personalized medicine. However, when the statistics are released as they are, there is a risk of identifying individuals. In this study, we propose efficient a nd practical privacy-preserving methods using the concept of differential privacy for linkage analysis with a transmission disequilibrium test (TDT). We focus on the case of two affected children in one family, and present differentially private data sharing methods based on three statistics, which are the TDT statistic, haplotype-based statistic, and combined statistic of these two. First, we show the sensitivities of each statistic and present the algorithm using the Laplace mechanism. Then, for the exponential mechanism, we adopt the shortest Hamming distance score as the score function and propose exact and approximation algorithms to find the scores. In our experiments, we measure the run time of each algorithm to show that it is feasible even on a large dataset containing 106SNPs. Supplementary materials are available at https://github.com/ay0408/DP-linkage-analysis-TDT.