Classifying contaminated cell cultures using time series features.

Area of Science:

• Cell Biology
• Biotechnology
• Analytical Chemistry

Background:

• Mycoplasma contamination is a common issue in mammalian cell cultures, affecting experimental results.
• Electric Cell-substrate Impedance Sensing (ECIS) provides real-time monitoring of cell behavior.
• Accurate and efficient detection methods for mycoplasma contamination are essential.

Purpose of the Study:

• To investigate the utility of Electric Cell-substrate Impedance Sensing (ECIS) time series data for differentiating mycoplasma-infected cell cultures from standard ones.
• To develop a classification method using low-dimensional features extracted from ECIS data for easy interpretation.
• To explore methods for mitigating experimental variations in ECIS measurements.

Main Methods:

• Extraction of application-relevant features from ECIS time course data.
• Implementation of low-dimensional feature-based classification for distinguishing between healthy and contaminated cell cultures.
• Analysis of experimental variations across different plates.

Main Results:

• High classification accuracy was achieved using only two specific features, dependent on the cell line.
• Initial findings indicate significant experimental variation between plates.
• Identification of feature types potentially more robust to plate-to-plate variation.

Conclusions:

• ECIS time series analysis combined with feature-based classification is a highly accurate method for detecting mycoplasma contamination.
• The study pioneers a broad examination of ECIS features for contamination detection.
• The research offers insights into managing and ameliorating experimental variations in ECIS assays.