An Efficient High-Throughput Screening of High Gentamicin-Producing Mutants Based on Titer Determination Using an Integrated Computer-Aided Vision Technology and Machine Learning.

Abstract

The "design-build-test-learn" (DBTL) cycle has been adopted in rational high-throughput screening to obtain high-yield industrial strains. However, the mismatch between build and test slows the DBTL cycle due to the lack of high-throughput analytical technologies. In this study, a highly efficient, accurate, and noninvasive detection method of gentamicin (GM) was developed, which can provide timely feedback for the high-throughput screening of high-yield strains. First, a self-made tool was established to obtain data sets in 24-well plates based on the color of the cells. Subsequently, the random forest (RF) algorithm was found to have the highest prediction accuracy with an R2 value of 0.98430 for the same batch. Finally, a stable genetically high-yield strain (998 U/mL) was successfully screened out from 3005 mutants, which was verified to improve the titer by 72.7% in a 5 L bioreactor. Moreover, the verified new data sets were updated on the model database in order to improve the learning ability of the DBTL cycle.