Fig 3. Method for identifying recurrent genetic part variants that likely arose from convergent evolution or engineering.

All plasmids containing the same genetic part variant are analyzed as a set. Segments shared by each pair of these plasmids are identified and queried against the full plasmid database. The results are used to calculate a design similarity (DS) score between the two plasmids. DS scores for all comparisons are used to construct a network graph of plasmid relatedness. Each separate cluster in the final graph is predicted to represent a set of plasmids in which the variant arose independently.