DNA Virus Detection System Based on RPA-CRISPR/Cas12a-SPM and Deep Learning

Our protocol introduces a rapid and highly sensitive detection system for frog virus 3. Significantly, this method enables point of care detection of DNA viruses, playing a crucial role in preventing the occurrence of pandemic diseases. The primary advantage of this technique lies in its integration of RPA with the CRISPR/Cas12a system, complimented by a portable smartphone-based microscope and AI classification.

This integration significantly enhances detection efficiency and accuracy while simultaneously reducing errors attributed to human factors. To begin, add the four key RPA enzymes in the reaction buffer. Then add the pre-designed primers to the mixture.

Vortex the mixture thoroughly. Add one microliter of the target DNA obtained from frog virus 3 to each RPA reaction, and vortex to mix well. Next, pipette seven microliters of 100 millimolar magnesium chloride to initiate the reaction.

Incubate the mixture at 37 degrees Celsius for 30 minutes to complete the assay. Prepare the Lachnospiraceae bacterium CAS12a protein with CRISPR RNA to form functional complexes. Mix the bacterial protein and 10 X CRISPR/Cas12a reaction buffer.

Now add 500 nanomolar single-stranded DNA reporter probe. Add one microliter of RPA reaction product to the CRISPR/Cas12a CRISPR RNA reaction mixture. Incubate the mixture at 37 degrees Celsius for 30 minutes.

Measure the fluorescent signals with a microplate reader. For detection with a smartphone microscope, first, pipette the prepared reaction mixture onto a retreated glass slide. Then cover it with a coverslip before incubation.

Place the slide on the stage of the smartphone microscope, and adjust the focal length and clarity. Capture an image to measure the fluorescent signals. Use ImageJ to measure the mean gray value of each image and the standard deviation of the mean gray value in a concentration group.

Adopt the deep learning model AlexNet 33 for classification. Now use Python to reshape the input images to 224 by 224 by three channels. Finally, use a pre-trained backbone network with ImageNet dataset to extract image features.

The sixth pair of primers provided the maximum amplification efficiency, and were selected. CRISPR RNA-3 proved to be the most efficient for collateral cleavage. The use of a developed detection system with selected RPA primers and CRISPR RNA resulted in significant differences between 10 aM FV3 and control.

The most important point to remember is the specific CAS 12a detection step. The CRISPR RNA of the reaction can be modified or redesigned to target other DNA viruses based on CAS 12a detection. The proposed method can assist in the preliminary assessment of the target virus's quantity.

Based on this, other methods like qPCR can be employed to obtain a more accurate viral load. This technique provides a preliminary attempt towards portable detection of DNA viruses. As for the frog virus 3 used in this protocol, timely detection can prompt effective protective measures, reducing losses in the breeding industry.