Integrating Microfluidics and Deep Learning to Investigate Entomopathogenic Nematode Responses to Host Cues

Area of Science:

• Biotechnology
• Entomology
• Bioengineering

Background:

• Entomopathogenic nematodes (EPNs) are vital biological control agents, offering an eco-friendly alternative to chemical pesticides.
• Optimizing EPN application requires a deep understanding of their physiological and behavioral responses to environmental cues.

Purpose of the Study:

• To investigate the behavior of Steinernema carpocapsae entomopathogenic nematodes (EPNs) in response to host-borne stimuli.
• To develop and validate a hybrid approach integrating microfluidics, deep learning, and optical flow for analyzing EPN behavior.

Main Methods:

• A microfluidic arena was designed to present stimuli to EPNs.
• A Convolutional Neural Network (CNN) model was employed to detect and analyze EPNs within the microfluidic system.
• Optical flow analysis was integrated to quantify EPN motor activity and dynamic responses.

Main Results:

• The CNN model accurately discriminated EPN behavior in the presence of host-borne cues, achieving high precision (1) and an F1-score of 0.933.
• Optical flow analysis revealed significantly increased motor activity in EPNs when exposed to stimuli.
• The hybrid approach successfully identified distinct motor behaviors associated with host cue detection.

Conclusions:

• This study provides novel insights into the ethology of Steinernema carpocapsae, demonstrating their heightened activity in response to host stimuli.
• The developed hybrid methodology offers a powerful tool for precise detection and comprehension of EPN responses.
• Findings advance the development of targeted and effective biocontrol strategies utilizing EPNs.