Overview:

While detecting COPD diseases, our team performed RCA (Rolling Cycle Amplification). However, in the real world, detection often requires a PCR machine or a Thermal Cycler. Even in using RCA, up to two devices are used to detect the existence of the disease or virus. If we wanted to promote early detection, our objectives were simple: a device or machine that could detect disease at a low cost was vital as a replacement. The significant challenges encountered while developing this device were effectively lowering the price and guaranteeing the test accuracy simultaneously. This required developing and designing the device using entirely different approaches and methods. This leads to extensive testing and debugging over the development process. The device is named VBD3 (Visual Based Disease Detection Device). By utilizing two ESP32 microcontrollers and sets of other electric components, a cost-efficient design, this device can also diagnose diseases and quantify the data for further analysis from doctors or lab members.

Parts List:

Design:

The electric system of VBD3 is built under an Arduino environment using Arduino IDE. However, Arduino itself is not a very powerful development board; there are a lot of other alternatives, including STM32, Raspberry Pi, Teensy boards…. etc. In our case, we used ESP32 development boards with the benefit of wireless communication, more substantial processing power, cost-efficiency…… etc. The electric system is mainly split into two parts: the heating and detecting mechanisms.

The heating mechanism is for heating the samples to a preferable temperature for them to react and perform RCA better; it uses a TEC1-12706 semi-conduct cooling pallet, a heating block, and a K-type thermocouple with MAX6675 module driving it for temperature sensing. These components are picked based on their reliability, sensitivity, and ability to heat or cool down components. The detecting system uses an ESP32-Cam module to take photos and a 12V blue LED to excite the fluorescence. The device also has an OLED screen, encoder, SD card module buzzers, and fans for controlling and cooling purposes.

The detecting mechanism of this device is straightforward. The sample is put into an aluminum holder and a heat block. With the program started, the heating system will begin heating the sample to a preset temperature, in our case, 37 Celsius. The heating mechanism will sustain the temperature through PID control, minimizing the inconsistency of the temperature. After this, the ESP32-cam module will take a photo of the sample consistently. However, right before it takes the picture, it uses a Blue LED light to illuminate the sample. After the photo is taken, the ESP32-cam sends the image to the ESP32 development board through the ESP NOW wireless communication protocol. When the ESP32 board successfully receives the photo, it performs the MGV (Mean Grey Value), where a captured picture of the sample will be converted to its Mean Grey Value. It saves the data into an SD card after taking photos for 2 hours and converting them to data, the device can export the data via an SD card to a mobile device or desktop, and the data is converted to a 2-axis graph for better-examining purposes.

Picture to MGV Conversion:

The conversion from picture to Mean Grey Value requires two steps. Including opening the image, and computing the Mean Grey Value. First, opening the image can be done easily by using a directory or libraries to read SD card information. The second step comes in a little more complexity, to compute the Mean Grey Value the following formula can be used;

N = the total number of pixels in the image (Width * Height)

To test the achievability of this method we performed the conversion of our sample picture to its MGV, with one of our samples being a sample that has fluorescence named Tpos and the other without fluorescence named Tneg (Figure 1).

Tpos:

Tneg:

By using the formula we get two MGV data one for each picture.

Based on the formula, we can tell that a higher MGV value then it usually means that there is more fluorescence in the sample, which also means that it is diagnosed positive for COPD. The result shows that the sample with fluorescence had an MGV of around 182 and the one without fluorescence had an MGV of around 174. This is well enough to distinguish the difference between positive and negative samples through a single greyscale picture of the sample.

Schematics and Images:

Below is a diagram (Figure 3) showing the detection method VBD3 uses for estimating the sample's fluorescence value. A blue LED light excites the sample and the emitted fluorescence light is captured by an ESP32-Cam camera through the filtration of lights through an optical filter.

As stated above (Figure 4) the electric system for VBD3 is split into two parts detecting and heating, the detecting part is what the MPU (Main Processing Unit) is in charge of, and the heating part is what the SPU (Secondary Processing Unit) is in charge of. The Microcontroller in MPU is the ESP32-WROOM-32U and is in charge of converting the images into values or MGV (Mean Grey Values). The other Microcontroller in SPU is the ESP32-Cam, which is in charge of capturing photos and PID controls of the heating mechanism. A brief electric diagram is shown in the image below (Figure 5-7).

Hardware Software:

The software is separated into two parts, one part for SPU and another part for MPU. The SPU is in charge of the code that heats the sample and sustains the heat through PID control, for capturing and exciting the sample. The MPU is in charge of the code that converts sample images into its MGV, exporting the data into a microSD card, displaying the status through an OLED screen, and controlling the device through an encoder.

Disease Diagnostic (Software):

After data collection through the detection device, the data can be exported via an SD card for further analysis and for diagnosing COPD. The software first allows to visualization of the data in a 2-axis graph. The main 3 functions of the software are to diagnose COPD, to diagnose the severity of the disease the patient has, and to remind patients to consume medicine on time. Based on our history interview, professional medical practitioners suggest that it is vital to detect COPD in the early stage and to follow the doctor’s advice to consume the medicine on schedule. Our device is capable of supporting potential patients to easily get access to COPD detection and our software can accurately diagnose COPD based on the change in fluorescence value of the data. Also, based on the concentration of the fluorescence our software can also tell how severe is the disease in the patient’s body. Through a systematic backup of data, our software is also capable of telling whether or not the patient’s condition is changing negatively or positively giving out further suggestions for mitigating the disease.

The code for software:

https://gitlab.igem.org/2023/software-tools/hs-china

Advantages:

1. An integrated experimental system combined with automated equipment means testing requires little to no involvement from professionals.

2. Optimizing the cost and size of the test machine greatly suits decentralized healthcare, such as community medical settings.

3. In a laboratory environment, the average testing duration is 30-60 minutes, and the precision of sample solution quantitative testing reaches the 20pmol level.

4. The biochemical system and hardware system are highly expandable, suitable for detecting biomarkers of other diseases.

References：

Jin, J., Vaud, S., Zhelkovsky, A., Pósfai, J., & McReynolds, L. A. (2016). Sensitive and specific miRNA detection method using SplintR Ligase. Nucleic Acids Research, 44(13), e116. https://doi.org/10.1093/nar/gkw399

He, Y., Wen, Y., Tian, Z., Hart, N. T., Han, S., Hughes, S. J., & Zeng, Y. (2023b). A one-pot isothermal Cas12-based assay for the sensitive detection of microRNAs. Nature Biomedical Engineering. https://doi.org/10.1038/s41551-023-01033-1