Simulation and machine learning modelling based comparative study of InAlGaN and AlGaN high electron mobility transistors for the detection of HER-2
| Autor: | Shivanshu Mishra, Nidhi Chaturvedi |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Receptor
ErbB-2 General Chemical Engineering chemistry.chemical_element Electrons Gallium 02 engineering and technology High-electron-mobility transistor Machine learning computer.software_genre Indium Analytical Chemistry law.invention Machine Learning 03 medical and health sciences 0302 clinical medicine law Humans Sensitivity (control systems) Aluminum Compounds High electron Human Epidermal Growth Factor Receptor 2 Training set business.industry Transistor General Engineering 021001 nanoscience & nanotechnology Aluminum gallium nitride chemistry 030220 oncology & carcinogenesis Artificial intelligence 0210 nano-technology business computer |
| Zdroj: | Analytical Methods. 13:3659-3666 |
| ISSN: | 1759-9679 1759-9660 |
| DOI: | 10.1039/d1ay00707f |
| Popis: | The detection of the cancer biomarker human epidermal growth factor receptor 2 (HER-2) has always been challenging at the early stages of cancer due to its very small presence. A systematic study of biosensors to achieve optimum sensitivity is of paramount significance. Thus, in this paper, we report a simulation study and machine learning (ML) based model for the comparative analysis of indium aluminum gallium nitride (InAlGaN) and aluminum gallium nitride (AlGaN) based high electron mobility transistors (HEMTs) for the detection of HER-2. The sensing performance of the InAlGaN based HEMT exhibits 1.8 times higher sensitivity as compared to that of the AlGaN based HEMT. The presented work also provides insights into the importance of the pH of the medium of HER-2. The results produced by the developed ML-based model are in good agreement with the simulation results. The model is not only capable of predicting within the trained range but also it can predict reasonably well even beyond the range of the training data. The introduction of a ML-based model significantly reduces the computational cost, time to perform similar type of simulations and, unlike the physics-based modelling, it also eliminates the need for empirical fitting of the model parameters. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|