Efficient Similarity Search on Encrypted Databases

在外包DNA数据库管理时，存储在数据库中或作为查询发送的序列信息应该受到数据库管理员的技术保护，因为它可能包含许多经济价值或隐私。在这项研究中，我们提出了一种基于Q-gram过滤的隐私保护DNA同源搜索方法，该方法通过对一组用于度量相似性的Q-gram进行扰动。此外，我们还证明了有间隙的Q-gram可以提高抗统计推断攻击的安全级别。

When outsourcing a DNA database administration, sequence information stored in the database or sent as a query should be technically protected from a database administrator since it may contain much economical values or privacy. In this research, we propose a privacy-preserving DNA homology search method with q-gram filtering by adding perturbation to a set of q-grams, which are used for measuring similarity. Also, we show that gapped q-gram can enchance the level of security against a statistical inference attack.