一种个人DNA数据隐私保护算法的改进

Improvement of a method of privacy protection for personal DNA data

DNALA是一种个人DNA数据隐私保护的方法。该方法能有效的实现对个人DNA数据的隐私保护,但前期数据预处理复杂,而且后期处理精度不高。本文针对DNALA的这些缺点进行改进,形成了Savior算法。Savior算法在数据预处理阶段用双序列比对代替了DNALA中的多序列比对,在随后的处理中用随机爬山法代替了DNALA中的贪心策略,从而克服了原算法的缺点。对比实验说明:在达到同样的保护强度时,Savior对数据的改动小于DNALA,数据预处理耗费的时间小于DNALA。

For the development of DNA sequencing technology and biology it is possible and needed to set up person- specific DNA databases. When setting up these databases, we must guarantee they are anonymous, that no one can find out whom a special DNA sequence is collected from. DNALA is a new method for doing this. Through overcoming the DNALA＇ s disadvantage - the pretreatment is complex and the precision of classing is not high enough, a new rnethod Savior is presented in this paper. Savior replaces the multiple alignment in DNALA with pairwise alignment between every tow sequences, and replaces the greedy algorithm in DNALA with stochastic hill- climbing. For doing this, it can overcome DNALA＇s shortcomings. The experiments show that when getting the same effect for protection, Savior use less time for pretreatment and gets higher precision than DNALA.