Early Prediction of COVID-19 Patient Survival by Targeted Plasma Multi-Omics and Machine Learning

Richard, Vincent; Gaither, Claudia; Popp, Robert; Chaplygina, Daria; Brzhozovskiy, Alexander G.; Kononikhin, Alexey; Mohammed, Yassene; Zahedi, René P.; Nikolaev, Evgeny N.; Borchers, Christoph H.

doi:10.1016/j.mcpro.2022.100277

Cited by 31 publications

(27 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several plasma proteomic datasets [ 9 – 15 , 17 , 19 , 20 ] were accumulated since the pandemic has started. We compared our dataset with five COVID‐19 plasma proteome studies [ 9 , 10 , 11 , 12 , 13 ] to find consistent alterations across different datasets (Figure 2A , Supporting Information File 3 ).…”

Section: Resultsmentioning

confidence: 99%

“…The impact of SARS‐CoV‐2 on human physiology can be monitored as plasma contains disease‐related biomarkers circulating in the blood that are released from different organs [ 16 ]. Plasma‐based proteomics studies compared the proteomic changes after SARS‐CoV‐2 infection between patients and controls from a single measurement [ 9 , 12 – 15 , 18 , 19 ] or with multiple measurements in a time‐dependent manner [ 10 , 11 , 17 , 20 ]. In general, inflammatory response, immune system‐related proteins [ 9 , 17 ], metabolic reconstitution, tissue repair‐related proteins, and regulators of coagulation [ 9 , 21 ] have been shown to be deregulated after infection [ 10 , 11 ].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Plasma proteomics identify potential severity biomarkers from COVID‐19 associated network

Sahin

Yurtseven

Dadmand

et al. 2022

Proteomics Clinical Apps

View full text Add to dashboard Cite

Purpose Coronavirus disease 2019 (COVID‐19) continues to threaten public health globally. Severe acute respiratory coronavirus type 2 (SARS‐CoV‐2) infection‐dependent alterations in the host cell signaling network may unveil potential target proteins and pathways for therapeutic strategies. In this study, we aim to define early severity biomarkers and monitor altered pathways in the course of SARS‐CoV‐2 infection. Experimental Design We systematically analyzed plasma proteomes of COVID‐19 patients from Turkey by using mass spectrometry. Different severity grades (moderate, severe, and critical) and periods of disease (early, inflammatory, and recovery) are monitored. Significant alterations in protein expressions are used to reconstruct the COVID‐19 associated network that was further extended to connect viral and host proteins. Results Across all COVID‐19 patients, 111 differentially expressed proteins were found, of which 28 proteins were unique to our study mainly enriching in immunoglobulin production. By monitoring different severity grades and periods of disease, CLEC3B, MST1, and ITIH2 were identified as potential early predictors of COVID‐19 severity. Most importantly, we extended the COVID‐19 associated network with viral proteins and showed the connectedness of viral proteins with human proteins. The most connected viral protein ORF8, which has a role in immune evasion, targets many host proteins tightly connected to the deregulated human plasma proteins. Conclusions and Clinical Relevance Plasma proteomes from critical patients are intrinsically clustered in a distinct group than severe and moderate patients. Importantly, we did not recover any grouping based on the infection period, suggesting their distinct proteome even in the recovery phase. The new potential early severity markers can be further studied for their value in the clinics to monitor COVID‐19 prognosis. Beyond the list of plasma proteins, our disease‐associated network unravels altered pathways, and the possible therapeutic targets in SARS‐CoV‐2 infection by connecting human and viral proteins. Follow‐up studies on the disease associated network that we propose here will be useful to determine molecular details of viral perturbation and to address how the infection affects human physiology.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Plasma proteomics identify potential severity biomarkers from COVID‐19 associated network

Sahin

Yurtseven

Dadmand

et al. 2022

Proteomics Clinical Apps

View full text Add to dashboard Cite

show abstract

“…AI has been successfully applied to biomarker identification based on MS data [10]. In a recent paper [66], machine learning was used to accurately predict the survival of COVID patients on the day they were admitted to the hospital ICU, based on the plasma concentrations of 10 proteins and 5 metabolites, as determined by MS. While humans can visually detect differences in the concentration of a single protein biomarker, but for sorting out the data from thousands of proteins, computerized methods are needed.…”

Section: Biomarker Discovery and Validationmentioning

confidence: 99%

“…Upper row -Proteomics model based on 10 proteins, middle row -Metabolomics model based on 5 metabolites, bottom row -Combined multi-omics model based on 10 proteins and 5 metabolites. (figure and figure legend from[66], with permission).…”

mentioning

confidence: 99%

Perspective Chapter: Pattern Recognition for Mass-Spectrometry-Based Proteomics

Bangert¹,

Balasubramaniam²,

Parker³

et al. 2024

Biomedical Engineering

View full text Add to dashboard Cite

Multiomic analysis comprises genomics, proteomics, and metabolomics leads to meaningful insights but necessitates sifting through voluminous amounts of complex data. Proteomics in particular focuses on the end product of gene expression – i.e., proteins. The mass spectrometric approach has proven to be a workhorse for the qualitative and quantitative study of protein interactions as well as post-translational modifications (PTMs). A key component of mass spectrometry (MS) is spectral data analysis, which is complex and has many challenges as it involves identifying patterns across a multitude of spectra in combination with the meta-data related to the origin of the spectrum. Artificial Intelligence (AI) along with Machine Learning (ML), and Deep Learning (DL) algorithms have gained more attention lately for analyzing the complex spectral data to identify patterns and to create networks of value for biomarker discovery. In this chapter, we discuss the nature of MS proteomic data, the relevant AI methods, and demonstrate their applicability. We also show that AI can successfully identify biomarkers and aid in the diagnosis, prognosis, and treatment of specific diseases.

show abstract

“…After learning the positive and negative gene distributions and the interaction networks of the associated genes, the model was given data for any gene and, with it having an AUC of 0.88, it revealed the probability of the gene as a potential target for anticancer drugs. In [ 143 ], a method was developed for the early prediction of COVID-19 patient survival by combining plasma multi-omics and DL. The precise concentration of 100 proteins and metabolites in the plasma from hospitalized patients was determined, and it appeared distinctively different from that of the control, healthy patients, thus, indicating the difference between the non-surviving patients and the surviving patients.…”

Section: Omics Data and Deep Learningmentioning

confidence: 99%

Omics Data and Data Representations for Deep Learning-Based Predictive Modeling

Tsimenidis

Vrochidou

Papakostas

2022

IJMS

View full text Add to dashboard Cite

Medical discoveries mainly depend on the capability to process and analyze biological datasets, which inundate the scientific community and are still expanding as the cost of next-generation sequencing technologies is decreasing. Deep learning (DL) is a viable method to exploit this massive data stream since it has advanced quickly with there being successive innovations. However, an obstacle to scientific progress emerges: the difficulty of applying DL to biology, and this because both fields are evolving at a breakneck pace, thus making it hard for an individual to occupy the front lines of both of them. This paper aims to bridge the gap and help computer scientists bring their valuable expertise into the life sciences. This work provides an overview of the most common types of biological data and data representations that are used to train DL models, with additional information on the models themselves and the various tasks that are being tackled. This is the essential information a DL expert with no background in biology needs in order to participate in DL-based research projects in biomedicine, biotechnology, and drug discovery. Alternatively, this study could be also useful to researchers in biology to understand and utilize the power of DL to gain better insights into and extract important information from the omics data.

show abstract

Early Prediction of COVID-19 Patient Survival by Targeted Plasma Multi-Omics and Machine Learning

Cited by 31 publications

References 66 publications

Plasma proteomics identify potential severity biomarkers from COVID‐19 associated network

Plasma proteomics identify potential severity biomarkers from COVID‐19 associated network

Perspective Chapter: Pattern Recognition for Mass-Spectrometry-Based Proteomics

Omics Data and Data Representations for Deep Learning-Based Predictive Modeling

Contact Info

Product

Resources

About