Quantifying Complexity in Metabolic Engineering using the LASER Database
Metabolic Engineering Communications, (July 2016)DOI: 10.1016/j.meteno.2016.07.002
Abstract
We present an updated LASER database for metabolic engineering analysis. A total of 433 E. coli and 190 S. cerevisiae designs are curated in this version. Complexity metrics describing difficulty and time for construction estimates are presented. Uses for these metrics to minimize labor, time, and cost of metabolic engineering projects are explored. We previously introduced the LASER database (Learning Assisted Strain EngineeRing, https://bitbucket.org/jdwinkler/laser\_release) to serve as a platform for understanding past and present metabolic engineering practices. Over the past year, LASER has been expanded by 50\% to include over 600 engineered strains from 450 papers, including their growth conditions, genetic modifications, and other information in an easily searchable format. Here, we present the results of our efforts to use LASER as a means for defining the complexity of a metabolic engineering ” design”. We evaluate two complexity metrics based on the concepts of construction difficulty and novelty. No correlation is observed between expected product yield and complexity, allowing minimization of complexity without a performance trade-off. We envision the use of such complexity metrics to filter and prioritize designs prior to implementation of metabolic engineering efforts, thereby potentially reducing the time, labor, and expenses of large-scale projects. Possible future developments based on an expanding LASER database are then discussed.