Identification of Synthetic Engineering in Prokaryotic Genomes Using Evolved Neural Networks

Synthetic biology affords the opportunity to engineer biological systems for enhanced behaviors of interest to humanity. However, along with this benefit comes with need to detect when engineering has occurred, for instance in the monitoring of foods said to be non-genetically modified, or to determine if a novel human pathogen was the result of natural evolution or synthetic engineering. Such decisions need to be made timely, repeatedly, at low cost, and in a manner that does not solely rely on organism phenotype. While it is possible to detect changes at the genetic level through genome sequencing, such processes are time intensive and costly. Here we demonstrate a proof-of-concept approach towards the detection of engineering in prokaryotic systems using neural networks optimized via evolutionary computation in light of knowledge of simulated edits and the natural diversity of the prokaryote Bacillus subtilis. The promising results can be extended to other systems for detection of synthetic engineering.