This study investigates the potential of bright-field (BF) microscopy images for discriminating single-gene mutant cells from wild-type cells using a machine learning approach. Hundreds of BF images of single-gene mutant cells were acquired, and single-cell profiles consisting of texture features were quantified. A machine learning model successfully discriminated mutants from wild-type cells (AUC = 0.773). The features contributing to this discrimination were identified and related to the morphology of intracellular structures. Functionally close gene pairs showed similar feature profiles in mutant cells, suggesting that BF image-based profiling holds potential as a tool for mutant cell profiling.

Godai Suzuki, Yutaka Saito, Motoaki Seki, Daniel Evans-Yamamoto, Mikiko Negishi, Kentaro Kakoi, Hiroki Kawai, Christian R. Landry, Nozomu Yachie, Toutai Mitsuyama