AI-assisted microfluidic immunoassay chip enabling early multiplex viral antibody detection in epidemics

Area of Science:

• Biotechnology
• Bioengineering
• Immunology

Background:

• Timely immune profiling is critical during viral outbreaks for effective containment and treatment.
• Conventional serological assays are often slow, require specialized equipment, and are not suitable for field deployment.

Purpose of the Study:

• To develop a rapid, field-deployable microfluidic immunoassay chip for viral diagnostics.
• To integrate artificial intelligence (AI)-guided epitope prediction with a novel microfluidic chip design.

Main Methods:

• Developed an AI model (ABEpre) for epitope prediction, achieving AUCs of 0.80-0.85.
• Created a searchable viral epitope database from AI predictions.
• Designed a pump-free polydimethylsiloxane (PDMS) microfluidic chip with micropillar arrays for enhanced antigen-antibody interactions.
• Utilized a manual dispenser for reagent introduction and indirect ELISA for visualization.

Main Results:

• Validated the chip with peptides from SARS-CoV-2, hepatitis B virus, and dengue virus.
• Demonstrated reproducible multiplex detection on-chip within 30 minutes.
• Confirmed antibody binding via ELISA.

Conclusions:

• The developed system is a structurally simple, low-cost platform for rapid outbreak response and point-of-care diagnostics.
• The study presents a fast sequence-to-chip pipeline linking epitope prediction to a deployable diagnostic tool for viral surveillance.