Security study of keyed DNA data embedding

The emergent field of DNA data embedding has, in the past decade, seen the proposal of many methods to encode data in the genome of living organisms. These algorithms have primarily focused on incorporating error correction into encoded data. However the security aspects of data encoded in DNA have generally been overlooked. As with any data hiding method, information encoded in DNA can be protected with a secret key known only to authorized individuals. While there have been suggestions to incorporate keys, no well defined functions for mapping keys to parameters have been proposed yet. Furthermore there have been no security analyses of keyed DNA data embedding methods, in terms of quantifying how difficult it is for an attacker to find the key. In this paper, modifications to incorporate secret keys into a prominent embedding algorithm are proposed, and the security is analysed under realistic attack scenarios. The algorithm to be modified, called BioCode ncDNA has been proposed by us, and is designed for encoding data in vivo (in the cells of living organism) with high biocompatibility and near-optimum embedding rate.