Glycomics meets artificial intelligence – Potential of glycan analysis for identification of seropositive and seronegative rheumatoid arthritis patients revealed

In this study, one hundred serum samples from healthy people and patients with rheumatoid arthritis (RA) were analyzed. Standard immunoassays for detection of 10 different RA markers and analysis of glycan markers on antibodies in 10 different assay formats with several lectins were applied for each...

Full description

Saved in:
Bibliographic Details
Published in:Clinica chimica acta 2018-06, Vol.481, p.49-55
Main Authors: Chocholova, Erika, Bertok, Tomas, Jane, Eduard, Lorencova, Lenka, Holazova, Alena, Belicka, Ludmila, Belicky, Stefan, Mislovicova, Danica, Vikartovska, Alica, Imrich, Richard, Kasak, Peter, Tkac, Jan
Format: Article
Language:English
Subjects:
Biomarker
Feedforward artificial neural network
Glycan
Glycoprotein
Immunoassay
Lectin
Machine learning algorithm
Rheumatoid arthritis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this study, one hundred serum samples from healthy people and patients with rheumatoid arthritis (RA) were analyzed. Standard immunoassays for detection of 10 different RA markers and analysis of glycan markers on antibodies in 10 different assay formats with several lectins were applied for each serum sample. A dataset containing 2000 data points was data mined using artificial neural networks (ANN). We identified key RA markers, which can discriminate between healthy people and seropositive RA patients (serum containing autoantibodies) with accuracy of 83.3%. Combination of RA markers with glycan analysis provided much better discrimination accuracy of 92.5%. Immunoassays completely failed to identify seronegative RA patients (serum not containing autoantibodies), while glycan analysis correctly identified 43.8% of these patients. Further, we revealed other critical parameters for successful glycan analysis such as type of a sample, format of analysis and orientation of captured antibodies for glycan analysis. •100 serum samples from healthy people and patients with rheumatoid arthritis (Ra) were analyzed.•Standard immunoassays and glycan detection based on application of lectins were performed.•A dataset containing 2000 data points was data mined using artificial neural networks.•Glycan analysis helped to correctly identify RA patients from healthy individuals.
ISSN:0009-8981
1873-3492
DOI:10.1016/j.cca.2018.02.031