Genomic Privacy Preserving Framework for High-Order SNPs Linkage Disequilibrium on Correlated Sequences

Considering the correlated SNPs sequences, the associated study between the SNPs and diseases may lead to more privacy breaches for high-order SNPs linkage disequilibrium, because of sensitive information related to SNPs including individual identity, phenotype and kinship. The existing privacy preserving methods have not been applicable to privacy preserving of SNPs, since those methods without considering correlation of the genome sequences, without effectively computation and with leading to utility disaster. To this end, we proposed the genomic privacy preserving framework for high-order SNPs linkage disequilibrium on correlated sequences. Firstly, we constructed a matrix model for diploid genotype of SNPs, while computing high-order SNPs linkage disequilibrium and the number of correlated sequences, then achieved diploid genotype indistinguishable based on the definition of differential privacy using matrix addition, finally, we got diploid genotype permutation using module operation. So we got the matrix differential privacy framework, which guaranteed the privacy of individual sensitive information for high-order SNPs linkage disequilibrium on correlated sequences, while kept the tradeoff of genome data utility. Therefore, combining matrix model of SNPs and modulo operation, our framework built on the definition of differential privacy has a guidance role for designing privacy preserving approach under high-order SNPs linkage disequilibrium on correlated sequences, while facilitates to share genome data to GWAS. Furthermore, our framework can generalize to privacy preserving of other genome data applications.