The Coronavirus Disease 2019 (COVID-19) global pandemic has had a profound, lasting impact on the world's population. A key aspect to providing care for those with COVID-19 and checking its further spread is early and accurate diagnosis of infection, which has been generally done via methods for amplifying and detecting viral RNA molecules. Detection and quantitation of peptides using targeted mass spectrometry-based strategies has been proposed as an alternative diagnostic tool due to direct detection of molecular indicators from non-invasively collected samples as well as the potential for high-throughput analysis in a clinical setting; many studies have revealed the presence of viral peptides within easily accessed patient samples. However, evidence suggests that some viral peptides could serve as better indicators of COVID-19 infection status than others, due to potential misidentification of peptides derived from human host proteins, poor spectral quality, high limits of detection etc. In this study we have compiled a list of 636 peptides identified from Sudden Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) samples, including from in vitro and clinical sources. These datasets were rigorously analyzed using automated, Galaxy-based workflows containing tools such as PepQuery, BLAST-P, and the Multi-omic Visualization Platform as well as the open-source tools MetaTryp and Proteomics Data Viewer (PDV). Using PepQuery for confirming peptide spectrum matches, we were able to narrow down the 639-peptide possibilities to 87 peptides that were most robustly detected and specific to the SARS-CoV-2 virus. The specificity of these sequences to coronavirus taxa was confirmed using Unipept and BLAST-P. Through stringent p-value cutoff combined with manual verification of peptide spectrum match quality, 4 peptides derived from the nucleocapsid phosphoprotein and membrane protein were found to be most robustly detected across all cell culture and clinical samples, including those collected non-invasively. We propose that these peptides would be of the most value for clinical proteomics applications seeking to detect COVID-19 from patient samples. We also contend that samples harvested from the upper respiratory tract and oral cavity have the highest potential for diagnosis of SARS-CoV-2 infection from easily collected patient samples using mass spectrometry-based proteomics assays.

中文翻译：

---

对基于 MS 的 2019 冠状病毒病 (COVID-19) 临床诊断的最佳肽靶标进行严格评估

2019 年冠状病毒病 (COVID-19) 全球大流行对世界人口产生了深远、持久的影响。为 COVID-19 患者提供护理并检查其进一步传播的一个关键方面是对感染进行早期和准确的诊断，这通常是通过放大和检测病毒 RNA 分子的方法来完成的。由于从非侵入性收集的样品中直接检测分子指标以及在临床环境中进行高通量分析的潜力，使用基于靶向质谱的策略对肽进行检测和定量已被提议作为一种替代诊断工具；许多研究表明，容易获取的患者样本中存在病毒肽。然而，有证据表明，由于人类宿主蛋白衍生的肽可能被错误识别、光谱质量差、检测限高等，某些病毒肽可以比其他病毒肽更好地指示 COVID-19 感染状态。在这项研究中，我们编制了从突发急性呼吸综合征冠状病毒 2 (SARS-CoV-2) 样本（包括体外和临床来源）中鉴定出的 636 种肽的列表。这些数据集使用基于 Galaxy 的自动化工作流程进行了严格分析，其中包含 PepQuery、BLAST-P 和多组学可视化平台等工具以及开源工具 MetaTryp 和蛋白质组数据查看器 (PDV)。使用 PepQuery 确认肽谱匹配，我们能够将 639 种肽的可能性缩小到 87 种肽，这些肽是检测最可靠且针对 SARS-CoV-2 病毒的。使用 Unipept 和 BLAST-P 证实了这些序列对冠状病毒分类群的特异性。通过严格的 p 值截止值与肽谱匹配质量的手动验证相结合，发现来自核衣壳磷蛋白和膜蛋白的 4 种肽在所有细胞培养物和临床样本（包括非侵入性收集的样本）中得到最可靠的检测。我们认为这些肽对于寻求从患者样本中检测 COVID-19 的临床蛋白质组学应用最具价值。我们还认为，使用基于质谱的蛋白质组学检测轻松收集的患者样本，从上呼吸道和口腔采集的样本最有可能诊断 SARS-CoV-2 感染。