Artificial Intelligence-Based Portable Bioelectronics Platform for SARS-CoV-2 Diagnosis with Multi-nucleotide Probe Assay for Clinical Decisions
| Autor: | Vikrant Singh, Madhuri Taranikanti, Srinivas Maddur, Ranjana Singh, Swati Mohanty, Vallepu Mohan Sai, Swathi Suravaram, Aswin Kumar Mudunuru, Shiv Govind Singh, Imran Ahmed Siddiqui, Lakshmi Jyothi Tadi, Patta Supraja, Suryasnata Tripathy, Rajiv Bandaru, Ch Gajendranath Chowdary, Tushant Agrawal, Rohith Kumar Guntuka |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Bioelectronics
Chemistry business.industry Heuristic (computer science) Nucleotides SARS-CoV-2 Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) COVID-19 Context (language use) Sensitivity and Specificity Analytical Chemistry Data acquisition COVID-19 Testing Artificial Intelligence Embedded system Humans RNA Viral User interface Biochip business Mobile device |
| Zdroj: | Analytical chemistry. 93(45) |
| ISSN: | 1520-6882 |
| Popis: | In the context of the recent pandemic, the necessity of inexpensive and easily accessible rapid-test kits is well understood and need not be stressed further. In light of this, we report a multi-nucleotide probe-based diagnosis of SARS-CoV-2 using a bioelectronics platform, comprising low-cost chemiresistive biochips, a portable electronic readout, and an Android application for data acquisition with machine-learning-based decision making. The platform performs the desired diagnosis from standard nasopharyngeal and/or oral swabs (both on extracted and non-extracted RNA samples) without amplifying the viral load. Being a reverse transcription polymerase chain reaction-free hybridization assay, the proposed approach offers inexpensive, fast (time-to-result: ≤ 30 min), and early diagnosis, as opposed to most of the existing SARS-CoV-2 diagnosis protocols recommended by the WHO. For the extracted RNA samples, the assay accounts for 87 and 95.2% test accuracies, using a heuristic approach and a machine-learning-based classification method, respectively. In case of the non-extracted RNA samples, 95.6% decision accuracy is achieved using the heuristic approach, with the machine-learning-based best-fit model producing 100% accuracy. Furthermore, the availability of the handheld readout and the Android application-based simple user interface facilitates easy accessibility and portable applications. Besides, by eliminating viral RNA extraction from samples as a pre-requisite for specific detection, the proposed approach presents itself as an ideal candidate for point-of-care SARS-CoV-2 diagnosis. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|