Detection of SARS-CoV-2 in nasal swabs using MALDI-MS

Abstract

Detection of SARS-CoV-2 using RT-PCR and other advanced methods can achieve high accuracy. However, their application is limited in countries that lack sufficient resources to handle large-scale testing during the COVID-19 pandemic. Here, we describe a method to detect SARS-CoV-2 in nasal swabs using matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) and machine learning analysis. This approach uses equipment and expertise commonly found in clinical laboratories in developing countries. We obtained mass spectra from a total of 362 samples (211 SARS-CoV-2-positive and 151 negative by RT-PCR) without prior sample preparation from three different laboratories. We tested two feature selection methods and six machine learning approaches to identify the top performing analysis approaches and determine the accuracy of SARS-CoV-2 detection. The support vector machine model provided the highest accuracy (93.9%), with 7% false positives and 5% false negatives. Our results suggest that MALDI-MS and machine learning analysis can be used to reliably detect SARS-CoV-2 in nasal swab samples. The outbreak of coronavirus disease 2019 (COVID-19) is a crisis that affects rich and poor countries alike 1. Detection of SARS-CoV-2 in patient samples is a critical tool for monitoring spread of the disease, guiding therapeutic decisions and devising social distancing protocols 2. Detection assays based on RT-PCR are the most effective and sensitive method for diagnosis of SARS-CoV-2 infection and are used in laboratories around the world 3. However, some countries lack the laboratory resources and access to PCR kits to conduct testing at the required levels. Therefore, other reliable diagnostic techniques are needed. Most clinical diagnostic laboratories have MALDI-MS equipment, which is used to identify bacterial and fungal infections. We propose to leverage the ease-of-use and robustness of MALDI-MS pathogen identification for large-scale SARS-CoV-2 testing in developing countries. MALDI-MS-based assays rely on reference spectra of strains and bioinformatics for high-sensitivity and high-specificity species identification through proteomic profiling. This approach is well established and accepted in many countries for routine diagnostics of yeast and bacterial infections. However, no spectral libraries for SARS-CoV-2 identification using MALDI-MS are publicly available to our knowledge. We first acquired MALDI mass spectra of nasal swab samples that had been tested for SARS-CoV-2 by RT-PCR and analyzed them using machine learning (ML). In this experiment (Fig. 1a), a total of 362 samples (211 SARS-CoV-2-positive and 151 negative, unequivocally confirmed by PCR), which came from three different countries, Argentina (Lab 1), Chile (Lab 2) and Peru (Lab 3), were placed on the MALDI plate without prior sample purification.

Open Science Resources for the Mass Spectrometry-Based Analysis of SARS-CoV-2

“…Many MS-based SARS-CoV-2 studies employ underpowered and problematic statistical and machine learning strategies that risk overfitting and could lead to incorrect results. Some authors have correctly stated that their results should indeed only be considered preliminary [8,21,41,42,55,63,84,89,99,106,110,135,142], whereas others have not clearly specified this caveat [17,76,81,95,107]. Furthermore, many of these studies do not make their raw data, identification results, feature tables, and code available (supplementary table 1), making it impossible to independently assess the validity of the results or to test alternative analysis workflows that could reveal additional data patterns that the original authors did not uncover with their chosen analysis approach.…”

“…As such studies continue to emerge, it is important to provide essential details such as a comparison to the viral load detectability, the limit of detection, and the specificity when compared to other methods such as RT-PCR. Alternatively, studies that aim to detect SARS-CoV-2 infection indirectly by measuring changes in host proteins and metabolites are also being introduced [63,81]. However, as in this case the viral proteins are not directly and unambiguously detected, MS signature-based diagnostics are an indirect readout compared to RT-PCR or targeted MS methods.…”

Proteomics in the COVID‐19 Battlefield: First Semester Check‐Up

“…[44] An adaptation of the method was developed, acquiring MALDI-TOF mass fingerprints for many nasal swab samples without prior sample purification. [45] In these data, the intensities of 88 peaks detected in the 3000-15500 m/z range were shown to discriminate between positive and negative SARS-CoV-2 samples but the methodology is not based on detection of SARS-CoV-2 proteins but rather host response, and has not been challenged against other infections. Interestingly, this approach had the same sensitivity as RT-qPCR in the most informative range (Cycle threshold values below 37).…”

A clinical MALDI-ToF Mass spectrometry assay for SARS-CoV-2: Rational design and multi-disciplinary team work

“…Consequently studies where swab samples have been split for simultaneous analysis by RT PCR detection systems of SARS-CoV-2 RNA and by MALDI-ToF mass spectrometry for viral proteins, are compromised [4].…”