This is a post-peer-review, pre-copyedit version of an article published in Multimedia Tools and Applications. The final authenticated version is available online at: https://doi.org/10.1007/s11042-019-08496-y

New genome sequencing technologies have simplified the generation of genomic data, making them more common but in turn a likely target of attack. Security strategies have been devised such as restricting the amount of information that can be queried or using new encryption techniques. These solutions might not be enough if the entire file has to be shared, as the recipient might leak the accessible information. This contribution addresses this issue using watermarking. Each read in a genomic file is modified depending on its content and a secret key. This allows generating different watermarked instances of the original file. Each watermark acts as a fingerprint: if a leak occurs, the unique modifications of the instance points to who originated the unauthorized publication. Using the key, the modifications can be undone. This allows sharing a leak-discouraging version with which the relevance of a file can be assessed, and can be reversed to the original if needed.

Peer Reviewed