The recent study published in Advanced Intelligent Systems, titled “The antimicrobial effect of plasma‐treated liquids (PTLs) arises from oxidative damage to microorganisms,” explores the promising intersection between machine learning and biomedical analysis. The research demonstrates how ML algorithms can predict the oxidative strength of plasma‐treated liquids, illustrating this capability through colorimetric assays. This approach offers a fresh perspective on optimizing antimicrobial treatments.
The study leverages machine learning (ML) and t‐distributed stochastic neighbor embedding (t‐SNE) to decode complex patterns in colorimetric images of cold atmospheric plasma (CAP)‐treated water. The research centers on CAP’s therapeutic potential, particularly focusing on reactive oxygen and nitrogen species (RONS) that contribute to antimicrobial efficacy.
ML and t-SNE in Analyzing RONS
Various color spaces, including RGB, HSV, LAB, YCrCb, and grayscale, are extracted from the colorimetric images representing oxidative stress caused by RONS. These features are then subjected to unsupervised machine learning using density‐based spatial clustering of applications with noise (DBSCAN). The evaluation of the DBSCAN model is based on metrics like homogeneity, completeness, and adjusted rand index, which are visualized using predictive data distribution graphs.
The study finds that the most effective results are obtained using the 3,3′,5,5′‐tetramethylbenzidine–potassium iodide colorimetric assay solution immediately after plasma treatment, yielding values of 0.894 for homogeneity, 0.996 for completeness, and 0.826 for adjusted rand index. To further refine these results, t‐SNE is employed for the best-case scenario to enhance clustering efficacy and identify optimal feature combinations.
Significance and Comparisons
Comparing this current study to previously published research, earlier work primarily focused on the qualitative analysis of colorimetric assays without the integration of sophisticated ML algorithms. Past methodologies lacked the precision and comprehensive analysis provided by the fusion of ML and t‐SNE, emphasizing the importance of modern computational techniques in enhancing biomedical research.
Additionally, earlier studies often relied on less dynamic approaches, which did not fully capture the intricate relationships present in the data. This study’s innovative use of ML and t‐SNE represents a significant advancement in the ability to analyze and visualize complex biomedical data, setting a new standard for future research in the field.
This integration of machine learning not only leads to more accurate predictions but also allows for a better understanding of the underlying mechanisms of CAP treatment. By utilizing advanced clustering techniques and visualization tools, the study offers a more detailed and comprehensive analysis of the oxidative strength of PTLs, which could enhance the development of more effective antimicrobial therapies. This approach also opens new avenues for the application of machine learning in various biomedical fields, potentially leading to more personalized and efficient treatments.
