On-demand microfluidic encapsulation of Caenorhabditis elegans in hydrogel for long-term observation via AI detection.

Biocompatible hydrogel encapsulation of model organisms, such as Caenorhabditis elegans, is highly desirable over chemical or mechanical immobilization methods that can harm health and hinder long term imaging studies. This study introduces an automated, gentle method for continuous immobilization within hydrogel blocks, achieved through flow lithography and machine learning-assisted vision detection. To optimize this automated worm immobilization, we investigated operating parameters, such as flow rate, solution composition, and worm loading concentration. With 90% precision in AI worm detection, we achieved 86% encapsulation efficiency at low to moderate worm-loading concentrations (1-7 worms per μL). Higher loading concentrations of 10-15 worms per μL reduced the encapsulation efficiency to 71%. Worm loading concentration, composition, and flow-rate ratios of the two input solutions-the worm-suspended aqueous solution (WSAS) and the polyethylene glycol diacrylate (PEGDA) solution (PEGS)-significantly affected worm throughput, worm dehydration during encapsulation, and the viability of encapsulated worms. The WSAS-to-PEGS flow-rate ratio was set at 0.5, with a minimum WSAS flow rate of 1 μL min-1 required to avoid backflow. The PEGDA concentration in PEGS was kept below 50% to avoid worm dehydration. By varying the concentrations of hydrogel precursor, PEGDA, we enabled both partial and complete immobilization suitable for short term and long-term studies, respectively. Results show the encapsulated worm's viability over 3 days, highlighting this method's potential for long-term studies without compromising worm health or movement.