Suhaimi Akmal, Zulkarnain Arash, Zin Noraziah Mohamad, Abdulhameed Abdullah, Kayani Aminuddin Ahmad, Buyong Ramdzan
Institute of Microengineering and Nanoelectronics (IMEN), Universiti Kebangsaan Malaysia, Bangi, 43600, Selangor, Malaysia.
Centre for Diagnostic, Therapeutic and Investigative Studies Faculty of Health Sciences Universiti Kebangsaan Malaysia Jalan Raja Muda Abdul Aziz, 50300, Kuala Lumpur, Malaysia.
Sci Rep. 2025 Jul 1;15(1):22079. doi: 10.1038/s41598-025-05274-9.
The gradual research in integrating artificial intelligence in the Dielectrophoresis system is rapid since the evolution of AI in every aspect of technology since the early 2020s. The benefits of AI integration into DEP systems include improving position and accuracy, faster processing and decision-making, enhancing particle classification, reducing human error, and many others. On the other hand, DEP research often focuses on CMF values of the particles. CMF values explain the behavior of the particle under the influence of the electric force in terms of trajectory and force magnitude. CMF values are calculated from the equation that requires the conductivity and permittivity of the medium and particles. One important aspect of CMF values is that they are non-numerical. Although possible, it is difficult to develop an algorithm using non-numerical values for detection applications. Hence, the study will focus on translating the non-numerical CMF values of E. coli and S. aureus into velocity (meters per second) and force (Newtons) parameters. In this study, we develop a simple method of calculating velocity and force units of bacterial movement using pixel coordinates and the time frame of the recorded video. From there, we managed to plot a force and velocity curve experimentally, with a crossover frequency of 1.0 to 1.5 MHz for S. aureus bacteria and 500 to 600 kHz for E. coli bacteria. Then, we validated our results with the velocity and force curves extracted from COMSOL simulation and the CMF curve extracted from MYDEP simulation. Our results show that the experimental curve plotted agrees with the simulation curve plotted from the COMSOL simulation, and the crossover frequency plotted in the experiment agrees with the CMF curve from MYDEP. The conclusion of the study is that the method developed in the study is important as the first step for the development of an artificial intelligence system to be integrated into the DEP system. The additional parameter of velocity alongside crossover frequency will improve the detection accuracy of bacterial cells using DEP technology. Furthermore, the collective data from future studies using this method will push DEP technology for future benefits.
自2020年代初人工智能在技术各方面取得进展以来，将人工智能集成到介电电泳系统中的相关研究进展迅速。将人工智能集成到介电电泳系统的好处包括提高定位和准确性、加快处理和决策速度、增强颗粒分类能力、减少人为误差等等。另一方面，介电电泳研究通常侧重于颗粒的CMF值。CMF值从轨迹和力大小方面解释了颗粒在电场力影响下的行为。CMF值由需要介质和颗粒的电导率和介电常数的方程计算得出。CMF值的一个重要方面是它们是非数值的。虽然有可能，但很难开发一种使用非数值进行检测应用的算法。因此，该研究将专注于把大肠杆菌和金黄色葡萄球菌的非数值CMF值转化为速度（米每秒）和力（牛顿）参数。在本研究中，我们开发了一种利用像素坐标和录制视频的时间帧来计算细菌运动速度和力单位的简单方法。据此，我们成功通过实验绘制了力和速度曲线，金黄色葡萄球菌的交叉频率为1.0至1.5兆赫兹，大肠杆菌的交叉频率为500至600千赫兹。然后，我们用从COMSOL模拟中提取的速度和力曲线以及从MYDEP模拟中提取的CMF曲线验证了我们的结果。我们的结果表明，绘制的实验曲线与从COMSOL模拟绘制的模拟曲线一致，实验中绘制的交叉频率与来自MYDEP的CMF曲线一致。该研究的结论是，本研究中开发的方法作为将人工智能系统集成到介电电泳系统的第一步很重要。除交叉频率外，速度这一附加参数将提高使用介电电泳技术检测细菌细胞的准确性。此外，未来使用该方法的研究收集的综合数据将推动介电电泳技术带来更多益处。
