Low false positive and accurate detection of yeast cell viability and concentration using an automatic staining and lensfree imaging platform

Abstract

Yeast cell viability and concentration are the crucial factors affecting product quality in food industry and bio-fuel production, as well as the evaluation basis for environmental toxic compounds. To overcome the drawbacks of existing methods, including high error, false positive and low automation, we propose a highly accurate approach based on an automatic staining and high-throughput lensfree imaging platform. A precisely controlled staining process is implemented automatically, which largely avoids the error caused by inappropriate exposure times. Based on optical simulation analysis, energy distribution characteristics are proposed. They are better with steady theoretical evidence for live yeast cell recognition. The parameters are directly extracted from raw cell fingerprints without any reconstruction. Those progresses improve robustness and increase efficiency. Availability of this approach is validated by compared the detection results with gold-standard PI counting method in a H2O2 toxicity test. So it is expected to be widely used in industrial production and environmental toxicity assessment.