Back

Defending Synthetic DNA Orders Against Splitting-Based Obfuscation

Tayouri, S.; Kogan, V.; Beal, J.; Levy, T.; Farbiash, D.; Flyangolts, K.; Mitchell, T.; Murphy, S. T.; Rotblat, B.; Veksler-Lublinsky, I.; Puzis, R.

2025-03-14 bioinformatics
10.1101/2025.03.12.642526 bioRxiv
Show abstract

Biosecurity screening of synthetic DNA orders is a key defense against malicious actors and careless enthusiasts producing dangerous pathogens or toxins. It is important to evaluate biosecurity screening tools for potential vulnerabilities and to work responsibly with providers to ensure that vulnerabilities can be patched before being publicly disclosed. Here, we consider a class of potential vulnerabilities in which a DNA sequence is obfuscated by splitting it into two or more fragments that can be readily joined via routine biological mechanisms such as restriction enzyme digestion or splicing. We evaluated this potential vulnerability by developing a test set of obfuscated sequences based on controlled venoms, sharing these materials with the biosecurity screening community, and collecting test results from open source and commercial biosecurity screening tools, as well as a novel Gene Edit Distance algorithm specifically designed to be robust against splitting-based obfuscations.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.