The Traveling Salesman Problem involves finding the shortest possible route visiting all destinations on a map only once before returning to the point of origin. The present study demonstrates a strategy for solving Traveling Salesman Problems using modified E. coli cells as processors for massively parallel computing. Sequential, combinatorial DNA assembly was used to generate routes, in the form of plasmids made up of marker genes, each representing a path between destinations, and short connecting linkers, each representing a given destination. Upon growth of the population of modified E. coli, phenotypic selection was used to eliminate invalid routes, and statistical analysis was performed to successfully identify the optimal solution. The strategy was successfully employed to solve a four-destination test problem.