Smartphone offers on-site soil microbiome profiling

The bacteria present in soil influence its nutrient cycling, plant growth, and overall health. Traditional techniques to characterize soil microbiomes, including polymerase chain reactions, gene sequencing, and biomarker analysis, often require specialized equipment as well as extensive time and training.

Liang et al. designed, built, and tested a portable, low-cost, and fast alternative. Their method integrates smartphone-based imaging with machine learning to classify bacterial species commonly found in soil samples and characterize soil health.

The authors capitalized on autofluorescence, the ability of some bacteria to fluoresce when excited by certain wavelengths of light. Previously, autofluorescence has been considered insufficient for differentiating bacterial species. The authors counteract this by using machine learning in their method.

The data is first collected by a smartphone equipped with an acrylic filter wheel to capture autofluorescence images of bacteria that have been extracted from soil and excited with light. Then, various algorithms, which were trained on large autofluorescence datasets, used the images to identify bacterial species and assess soil health.

The authors tested their method on individual bacteria species, various bacterial mixtures, and soil samples. The platform achieved an average accuracy of 88 percent in distinguishing dominant bacteria species. It was able to characterize soil health from simulated samples and, when tested on real soil samples, it achieved 80 percent overall accuracy.

“This study demonstrates the potential of this smartphone-based system as a valuable tool for on-site soil assessment, microbial monitoring, and environmental management,” said author Jeong-Yeol Yoon.

The authors are considering developing a smartphone app for this system that could be easily used in the field.

Source: “A smartphone-based approach for comprehensive soil microbiome profiling,” by Yan Liang, Bradley Khanthaphixay, Jocelyn Reynolds, Preston J. Leigh, Melissa L. Lim, and Jeong-Yeol Yoon, Applied Physics Reviews (2024). The article can be accessed at https://doi.org/10.1063/5.0174176 .

This paper is part of the Materials and Technologies for Bioimaging and Biosensing collection, learn more here .