Genetic similarity of biological samples to counter bio-hacking of DNA-sequencing functionality

We present the work towards strengthening the security of DNA-sequencing functionality of future bioinformatics systems against bio-computing attacks. Recent research has shown how using common tools, a perpetrator can synthesize biological material, which upon DNA-analysis opens a cyber-backdoor for the perpetrator to hijack control of a computational resource from the DNA-sequencing pipeline. As DNA analysis finds its way into practical everyday applications, the threat of bio-hacking increases. Our wetlab experiments establish that malicious DNA can be synthesized and inserted into E. coli, a common contaminant. Based on that, we propose a new attack, where a hacker to reach the target hides the DNA with malicious code on common surfaces (e.g., lab coat, bench, rubber glove). We demonstrated that the threat of bio-hacking can be mitigated using dedicated input control techniques similar to those used to counter conventional injection attacks. This article proposes to use genetic similarity of biological samples to identify material that has been generated for bio-hacking. We considered freely available genetic data from 506 mammary, lymphocyte and erythrocyte samples that have a bio-hacking code inserted. During the evaluation we were able to detect up to 95% of malicious DNAs confirming suitability of our method.