Related Experiment Video
Updated: Nov 11, 2025
07:20
Detecting the Lyme Disease Spirochete, Borrelia Burgdorferi, in Ticks Using Nested PCR
Published on: February 4, 2018
18.4K
Predicting Lyme Disease From Patients' Peripheral Blood Mononuclear Cells Profiled With RNA-Sequencing
Daniel J B Clarke1, Alison W Rebman2, Allison Bailey1
1Department of Pharmacological Sciences, Mount Sinai Center for Bioinformatics, Icahn School of Medicine at Mount Sinai, New York, NY, United States.
Frontiers in Immunology
|March 25, 2021
Summary
This study reveals that RNA sequencing can distinguish Lyme disease patients from controls, showing persistent immune gene changes over a year. Machine learning models identified Lyme disease but could not predict post-treatment persistent symptoms.
Area of Science:
- Immunology
- Genomics
- Computational Biology
Background:
- Lyme disease is a prevalent yet often under-diagnosed tick-borne illness.
- Understanding the molecular and cellular changes in Lyme disease is crucial for diagnosis and management.
Purpose of the Study:
- To investigate the longitudinal changes in gene expression and cell composition in patients with acute Lyme disease.
- To develop machine learning classifiers for distinguishing Lyme disease patients from controls and other conditions.
Main Methods:
- RNA sequencing of peripheral blood mononuclear cells from 73 acute Lyme disease patients and controls over one year.
- Extensive clinical phenotyping and deconvolution analysis to assess cell type composition.
- Development and evaluation of machine learning classifiers for disease classification.
Main Results:
- RNA sequencing data clearly separated Lyme disease cases from controls, with persistent alterations observed over time.
- Enrichment analysis identified significant up-regulation of immune response genes in Lyme disease patients.
- Machine learning models successfully distinguished Lyme disease patients from controls and COVID-19 patients.
- Classification models were unable to predict post-treatment persistent symptoms in early Lyme disease cases.
Conclusions:
- Gene expression profiling and machine learning offer potential tools for Lyme disease diagnosis.
- Lyme disease induces lasting changes in immune gene expression and cell composition.
- Predicting the development of post-treatment persistent symptoms remains a challenge.
