Deep-dLAMP: Deep Learning-Enabled Polydisperse Emulsion-Based Digital Loop-Mediated Isothermal Amplification.

Area of Science:

• Biotechnology
• Molecular Biology
• Artificial Intelligence

Background:

• Digital nucleic acid amplification tests offer absolute quantification but require costly specialized instruments, hindering widespread adoption.
• Existing methods for digital nucleic acid quantification often rely on uniform droplet generation and fluorescence detection, which are instrument-intensive.

Purpose of the Study:

• To develop a low-cost, label-free digital nucleic acid quantification method using deep learning and polydisperse emulsions.
• To enable absolute quantification of nucleic acids without expensive, specialized equipment.

Main Methods:

• Implemented loop-mediated isothermal amplification (LAMP) within polydisperse emulsions.
• Utilized a deep learning algorithm to analyze emulsion images, segment droplets, and determine occupancy status based on precipitated byproducts.
• Applied Poisson distribution to infer nucleic acid concentration from emulsion volume and occupancy data.

Main Results:

• Achieved accurate prediction of emulsion sizes and occupancy status.
• Demonstrated accurate nucleic acid concentration measurements with a limit of detection of 5.6 copies/µl and a dynamic range of 37.2 to 11000 copies/µl.
• Showcased robust performance across varying experimental parameters, including vortexing time and image quality.

Conclusions:

• Deep learning-enabled polydisperse emulsion-based digital loop-mediated isothermal amplification (deep-dLAMP) significantly reduces instrument costs for digital nucleic acid tests.
• The deep-dLAMP method offers a promising advancement for accessible nucleic acid quantification in biomedical laboratories and potential point-of-care diagnostics.