Identification of SARS-CoV-2 in nasal swabs with MALDI-MS and machine learning methods

AP19578688

Aim: To evaluate the rapid method for determining the SARS-CoV-2 virus in material obtained from a nasal smear using mass spectrometry and machine learning methods.

Relevance

Coronavirus infection, appeared in 2019, has become a challenge for humanity and for the health system in particular. The epidemic revealed many gaps in the organization of the diagnostic process, leading to the collapse of the laboratory service in the Republic of Kazakhstan during the periods of the highest incidence. To detect viral RNA in biomaterial, reverse transcription PCR (RT-PCR) is used. The detection of SARS-CoV-2 using RT-PCR tests is a highly accurate method, but the use of this diagnostic method is costly and time consuming.

Our research team is proposing to evaluate a method for detecting SARS-CoV-2 in nasal swabs using matrix laser desorption / ionization mass spectrometry (MALDI-MS) and machine learning analysis. This approach uses equipment and skills that are routinely used in clinical laboratories. The biomaterial does not require preliminary sample preparation and expensive reagents. Thus, the proposed method for the detection of SARS-CoV-2 in material obtained from a nasal swab will become a more affordable and less time-consuming test than RT-PCR.

Expected results

Based on the results of the scientific research, we evaluated previously existing methods and also developed our own express method for detecting SARS-CoV-2 in biomaterials obtained from nasal swabs. This new approach in COVID-19 diagnostics has a potential socio-economic impact, resulting in reduced diagnostic search time, timely adequate treatment of critical conditions, shortened hospitalization periods, and consequently, reduced treatment costs. The implementation of the research will help reduce mortality rates and increase life expectancy.

Obtaining information about new methods of early COVID-19 diagnosis will allow changes to be made to the diagnostic algorithm. This will be innovative not only within the Republic of Kazakhstan but also worldwide.

  1. Kadyrova I.A. - PhD in the specialty "Medicine", associate professor, doctor of clinical and laboratory diagnostics, senior researcher of the Research laboratory of NJSC «KMU». h-index - 5, ORCIDID 0000-0001-7173-3138.
  2. Babenko D.B. - PhD in the specialty "Medicine", specialist in molecular genetics research methods, bioinformatics, statistical analysis and machine learning methods. h-index - 5 (Scopus); 2 (Web of Knowledge), Author ID Scopus: 55935641200 SPIN-код: 3567-2768, Researcher ID P-8052-2017, ORCID 0000-0003-2280-2146.
  3. Egorov Sergey - PhD in "Immunology" (University of Toronto). In the project, he acts as a consultant on immunological methods of research and interpretation of immunopathogenesis. h-index – 17, ORCID ID 0000-0002-7136-7921.
  4. Lavrinenko A.V. - Master in the specialty "Medicine", doctoral student of 3 years of study, researcher of the Research laboratory of NJSC "KMU" of NJSC «KMU». h-index - 1, ORCID 0000-0001-9436-8778.
  5. Kolesnichenko S.I. - Junior Researcher of the Research laboratory of NJSC «KMU» . ORCID 0000-0003-3515-8900.
  6. Stupina E.A. - Head of the department of drip and neuroinfections, infectious disease doctor of the regional clinical hospital in Karaganda.
  7. Sultanbekova A.A. - otolaryngologist, PhD doctoral student.

The project implementation has been fully completed.

The research team evaluated the performance of the insilico machine learning algorithm on combined data from South America and Kazakhstan and also trained their own ML models with the aim of differentiating samples with SARS-CoV-2. The SVM-R and DT models showed especially high performance, surpassing other models, with ROCAUC scores of 0.983 and 0.972, respectively. SVM-R is particularly effective in distinguishing samples from Kazakhstan, achieving an accuracy of 88.0% for SARS-CoV-2+, 95.0% for NCARI, and 78.0% for AC. The potential of MALDI MS as an innovative diagnostic tool for identifying various respiratory infections is emphasized, as well as the need for further optimization and validation of the MALDI MS/ML method before its integration into routine clinical practice, considering the changing epidemiological situation.

The areas of application of the research results include: infectious diseases, clinical microbiology, laboratory diagnostics.

Научные публикации в рамках проекта:
    2021 year:
  1. Lavrinenko A., Kolesnichenko S., Turmukhambetova A., Kadyrova I. Respiratory pathogens co-infection in patients with COVID-19 pneumonia in Kazakhstan//Microbiology society. Annual conference online 2021. Poster abstract book 26–30 April 2021, Р 282.
    2. 2022 year:
  2. Kadyrova I, Yegorov S, Negmetzhanov B, Kolesnikova Y, Kolesnichenko S, Korshukov I, et al. (2022) High SARS-CoV-2 seroprevalence in Karaganda, Kazakhstan before the launch of COVID-19 vaccination. PLoS ONE 17(7): e0272008. https://doi.org/10.1371/journal. pone.0272008 (Scopus -87 процентиль, Q1).
  3. Kadyrova I.A., Sultanbekova A.A., Barkhanskaya V.I., Kolesnichenko S.I., Kolesnikova E.А., Lavrynenko А.V., Korshukov I.V., Bekov Ye.K., Yegorov S.V., Babenko D.B. Application of MALDI MS mass spectrometry for the detection of SARS-CoV-2 in nasopharyngeal swabs//Systematic review Science & Healthcare, 2022. (Vol. 24) 5, Р. 36-44. (КОКСОН).
  4. I. Kadyrova, S. Kolesnichenko, Ye. Kolesnikova, I. Korshukov, V. Barkhanskaya, A. Sultanbekova, D. Babenko SARS-COV-2 Detection in MALDI-TOF Mass Spectra by Machine Learning/ Int'l Conference on Mathematical Modeling in Physical Sciences September 5-8, 2022 Virtual, on-line Conference
  5. S.I. Kolesnichenko, I.A. Kadyrova, I.V. Korshukov, Ye.A. Kolesnikova, V.I. Barkhanskaya, D.B. Babenko SARS-COV-2 DETECTION USING MALDI-TOF AND MACHINE LEARNING APPROACH// IX International Conference of Young Scientists: Virologists, Biotechnologists, Biophysicists, Molecular Biologists and Bioinformaticians within the framework of the OPENBIO Open Communications Platform.
    6. 2023 year:
  6. Yegorov S., Kadyrova I. et al «Application of MALDI MS and Machine Learning to Detection of SARS-CoV-2 and non-SARS-CoV-2 Respiratory Infections» находится на рецензировании в журнале PLOS One, опубликован препринт https://doi.org/10.1101/2023.08.31.23294891
  7. А.А. Султанбекова, И.А. Кадырова, Д.Б. Бабенко. Maldi-Tof масс-спектрометрия жәнемашиналық оқыту әдістерін қолдануарқылы sars-cov-2 вирусындиагностикалаудың экспресс әдісін бағалау// Фармакология Казахстана, 2023, № 1, С. 262-276
  8. А.А. Султанбекова. Опыт применения время пролетной масс-спектрометрии (MALDI-TOF) для детекции вирусных патогенов, в частности короновируса. Систематический обзор Фармакология Казахстана, 2023, № 4, С. 97-106.
  9. Kadyrova, I., Kolesnichenko, S., Korshukov, I., Kolesnikova, Y., Barkhanskaya, V., Lavrinenko, A., Babenko, D. (2023). SG-APSIC1053: Detection of SARS-COV-2 in nasopharyngeal swags with MALDI-TOF MS and machine learning. Antimicrobial Stewardship & Healthcare Epidemiology, 3(S1), S2-S3. doi:10.1017/ash.2023.9
  10. Dmitriy Babenko et al Application of MALDI MS to Differentiate SARSCoV-2 and Non-SARS-CoV-2 Symptomatic Infections in the Early and Late Phases of the Pandemic// Conference Proceedings, Zurich Switzerland July 24-25, 2023.

    11. Patent and Implementation Acts.

    Kadyrova I, Kolesnichenko S, Korshukov I, Barkhanskaya V, Sultanbekova A, Kolesnikova E, Babenko D. Certificate of registration in the state registry of rights to objects protected by copyright No. 24868 dated April 7, 2022. Development of an express method for detecting the SARS-CoV-2 virus in material obtained from a nasal swab using Time-of-Flight Mass Spectrometry and machine learning methods.

    Two implementation acts of the research "Express method for detecting the SARS-CoV-2 virus in material taken from a nasal swab using Time-of-Flight Mass Spectrometry and machine learning methods" included in the list of methods of the Research Laboratory and in clinical practice of the Clinic of NСJSC «Medical University of Karaganda».