Team:NCTU Formosa/Hardware

To bring E. Phoenix to everyday life, we need to provide a workspace that enables our biology part to work safe and efficient. Therefore, Detector EP was born, a device that integrates sensors, incubating platform, and biosafety design. Besides, we construct a web page to showcase the results collected by Detector EP. Combining with our Mutagenicity Prediction AI, users can get all the information about the testing target with a few clicks on the Internet. Furthermore, we developed Visualizer EP that can turn mutagenicity results in a race car game, which makes our concept easy-understanding and fun. Behind all these parts is an IoT system that formed our system structure, and brought everything together.

IoT System

Figure 2: Device connected with IoT system

Introduction

By getting the mutagenicity of environment, our model can improve itself by keep updating the data and factors. Through our AI machine learning system combined with IoT(internet of things) system, Arduino Yun and sensor in the field, the information could be swiftly transported. As more data we collect from sensorbox, the mutagenicity prediction would work more accurately, giving user a direct, precise idea of the dangerousness of the chemical.

Procedure

1. Construct our IoT system with Arduino Yun and sensors box.

2. Arduino Yun receives the conditions and transmit to our cloud
server through wifi.

3. Current condition would be check on the website or server
terminal.

System Design

This is the schematic diagram on the design of our Arduino Yun. We connect, OD sensors, temparature sensor and power module with Arduino Yun. So that all the data from the box can be collected and analyzed.

Figure 3: The IoT system design

The data can be shown in the website clearly so that the users can real time monitor the device(which is not necessary). Most importantly, predictions results can also been shown after data analysis. By this system, the mutagenicity of the tested sample can be managed automatically and provide advices to users to achieve better safety awareness.

Detector EP

Figure 4: Detector EP final version

Introduction

Figure 5: Perspective schematic view of our device

Starting with our initial concepts that we want to raise public awareness of health and environment, we created a hardware biosensor called Detector EP to predict mutagenicity. In order to fully implement our system to the public, we designed Detector EP which can evaluate the danger outside the lab without needing any special training to operate the device. We used 3D printing as well as the open source Arduino system to form a prototype of the hardware, which is portable and IoT operative. Users could use the device to quickly detect and grasp a preliminary screening of unknown samples in the device and finally the data are transmitted to IoT platform online for data analysis, and the preliminary screening output, the mutagenicity, could be viewed on the website by the users.

Detector EP features its user-friendliness, semi-automation, and portability. The entire process startingfrom detecting unknown samples to giving a result of mutagenicity happens in the 22.6x13x20 centimeter device. Moreover, this device is capable of monitoring the instant and continuous condition in the detection box and garnering the measurements on our website platform. The following details about the Detector EP consists of three main parts: the mini incubator, the sensor, and biosafety

Figure 6: The Detector EP working process

The detection process of mutagenicity in the hardware system is based upon the logic of wet lab mutagenicity detection, but performed in the hardware sensor and analyzed on the IoT platform. For starter, the glass flasks contains 2 mLs of engineered E. coli, one of which is the control group for reference, and the other is the experiment group. The two flasks will be inserted on the shaker with lids with perforated 0.22 micrometer in diameter pores. The culturing process will be documented with O.D.sensors every 5 minutes and the recorded data will be uploaded on clouds in real-time. The users will be reminded to add IPTG and the unknown chemical of interest when the E. coli grows and the O.D. reaches 0.3. The following detected value by the O.D. seneor will also be sent to our website continually through IoTtalk. Secondly, the detected O.D. values will be transferred and analyzed on websites and computed into mutagenicity by growth curve model and eventually, show the final results on our website.

Mini Incubator

Figure 7: The incubator prototype

The mini incubator is designed to contain two tubes with engineered E. coli for incubation. It consists of three parts, the heater, the incubator wall and the shaker. We used PTC heater to heat the incubator, and the innerest layer of incubator wall is covered with copper to make the incubation atmosphere stable in 37 degrees celsius in uniform. Besides, the shaker has a traingular PLA pedestal with two cylnder hollows for incubation flasks, fixed with four springs at the two corners to make the shaker flexible and also stable in position during the shaking process.

Figure 8: (A) The shaker sketch and (b) the shaker prototype

The shaker consists of a motor and a bearing. The bearing increases its shaking diameters and the speed of shaking.

Sensor

Figure 9: The Sensor part

The sensor constitutes an Arduino board, and two sensors including an O.D. sensor and temperature sensor. The temperature sensor sends real-time temperature to the internet, monitoring the whole Detector EP and the sensor screens the absorbance of the bacterial fluid in the glass flasks. Once the concentration reaches O.D. 0.3, a notification will be sent to the users, reminding them of the timing to induce IPTG and add the chemical substance

Figure 10: The O.D. sensor illustration

It is necessary to confirm the accuracy in our sensor system. So we compared our measurement in the sensor box with the real value measured by Elisa plate reader. Our analysis results are shown in Figure 11.

Figure 11: Comparison between sensor readout and O.D. reader readout.

The resulting graphs show our measurements have a high positive correlation with the real value that our sensor system is accurate enough to measure the growth of E. coli in the Sensor Box.

Biosafety

Biosafety and biosecurity are of significant concern when working with possibly contaminated samples and engineering microorganisms. Engineering E. coli, which cultured in our device, may stronger than those wild types. It could harm the environment when it is released outside. To avoid the bacteria from surviving in nature in case an accidental spill, we created the security mechanism: double shields.

Double shields

We design double-shelded protection to prevent the leakage of Engineering E. coli. Bleaching water is carefully placed in the space between the two walls, and when the device is accidentally damaged, the bleaching water will overflow to eliminate the bacteria cultured in the device. The material we used to assemble the shell is polylactic acid (PLA), which can resist the corrosiveness of bleach and have enough density to keep the bleaching water in the folds.

Material Required

Table 1: The Detector EP budgets

Meterial Required	Quantity	Cost
3D printer material polylactide	2	26.3 USD
0.8mm copper plate	3	11 USD
PTC heater	1	2.83 USD
Thermometer DS18B20	1	3.36 USD
Light sensor TSL235R	2	7 USD
LED 525nm	2	1.67 USD
Motor	1	9.67 USD
Bearing	1	3.5 USD
Arduino YUN	1	73.33 USD
Relay	2	2.4 USD
Spring	8	2.83 USD
Total		143.89 USD

Visualizer EP

Let’s race!!!!!!!!!!!

The final goal of our project is to get most people be aware of the gene mutation, and show how different mutagenicity could be visualized.

“Visualizer EP” is then the product we developed. A car racing IoT game combined with mutagenicity predicting system. All users have to do is connect the IoT model cars to internet and get on to our web page with a searching system, typing the name of the chemical compound.

Steps for Playing the Racing Game:

1. Connect the model cars to the internet.

2. Type in the name of the chemical compound (type to if you want to race).

3. Press “Race”

4. Watch the racing game!

The model cars are carrying the Arduino Yun mini, an Arduino board with wifi connection. With the burned in code, it receives the signal from the IoTtalk.

Website

On the other hand, there is a mutagenicity predicting system on the website. As our dry lab team developed the data set, a python-based machine learning was built in an elegant way. This is how the magic happened:

1. Users type in the name of the compound.

2. The name of the compound will be transformed into SMILE form.

3. The SMILE file is sent in to the data set with the predicting formula.

4. The mutagenicity output will go through a linear transformation then be sent to IoTTalk.

5. The cars will receive the output and Runnnnnnn!

Figure 12: The Visualizer EP website

Reference

1. Lin, Y., et al. (2017). "IoTtalk: A Management Platform for Reconfigurable Sensor Devices." IEEE Internet of Things Journal 4(5): 1552-1562.

Go to >>>

Demonstration