Assessing the Protective Metabolome Using Machine Learning in World Trade Center Particulate Exposed Firefighters at Risk for Lung Injury - PubMed Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Sep 3;9(1):11939.
doi: 10.1038/s41598-019-48458-w.

Assessing the Protective Metabolome Using Machine Learning in World Trade Center Particulate Exposed Firefighters at Risk for Lung Injury

George Crowley  1 , Sophia Kwon  1 , Dean F Ostrofsky  1 , Emily A Clementi  1 , Syed Hissam Haider  1 , Erin J Caraher  1 , Rachel Lam  1 , David E St-Jules  2 , Mengling Liu  3   4 , David J Prezant  5   6 , Anna Nolan  7   8   9
Affiliations

Assessing the Protective Metabolome Using Machine Learning in World Trade Center Particulate Exposed Firefighters at Risk for Lung Injury

George Crowley et al. Sci Rep. .

Abstract

The metabolome of World Trade Center (WTC) particulate matter (PM) exposure has yet to be fully defined and may yield information that will further define bioactive pathways relevant to lung injury. A subset of Fire Department of New York firefighters demonstrated resistance to subsequent loss of lung function. We intend to characterize the metabolome of never smoking WTC-exposed firefighters, stratified by resistance to WTC-Lung Injury (WTC-LI) to determine metabolite pathways significant in subjects resistant to the loss of lung function. The global serum metabolome was determined in those resistant to WTC-LI and controls (n = 15 in each). Metabolites most important to class separation (top 5% by Random Forest (RF) of 594 qualified metabolites) included elevated amino acid and long-chain fatty acid metabolites, and reduced hexose monophosphate shunt metabolites in the resistant cohort. RF using the refined metabolic profile was able to classify cases and controls with an estimated success rate of 93.3%, and performed similarly upon cross-validation. Agglomerative hierarchical clustering identified potential influential pathways of resistance to the development of WTC-LI. These pathways represent potential therapeutic targets and warrant further research.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Study Design.
Figure 2
Figure 2
Demonstration of Model Optimization: PCA Scores Plot. (A) PCA of the qualified profile reveals heterogeneity in the data. (B) PCA of the refined profile demonstrates improved class separation produced by the refined profile compared to initial PCA (panel A).
Figure 3
Figure 3
(A) Qualified Profile PCA Loading Weights Plot. Loading weights plot of PCA of the qualified profile shows ill-defined metabolite clustering. (B) Refined Profile PCA Loading Weights Plot was used to derive insight into possible associations of biomarkers.
Figure 4
Figure 4
Random Forests Variable Importance in Projection. RF variable importance in projection is measured by mean decrease accuracy; the top 5% of metabolites important to class separation are shown. The confusion matrix shows classification accuracy of the refined profile.
Figure 5
Figure 5
Agglomerative, Hierarchical Clustering. (A) Data Matrix. Clustering of the data matrix identified 6 clusters of metabolites (A-F) and separated resistantWTC-LI from controls. (B) Correlation Matrix. Clustering of the correlation matrix reveals 4 clusters of metabolites (1-4) and intercluster correlations.
Figure 6
Figure 6
Pathway Schematics. Pathway schematics of fatty acid metabolism and the hexose monophosphate shunt. Node size correlates to fold change, red indicates fold change >1, resistantWTC-LI/control.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Naveed B, et al. Metabolic syndrome biomarkers predict lung function impairment: a nested case-control study. Am J Respir Crit Care Med. 2012;185:392–399. doi: 10.1164/rccm.201109-1672OC. - DOI - PMC - PubMed
    1. Friedman SM, et al. Case-control study of lung function in World Trade Center Health Registry area residents and workers. Am J Respir Crit Care Med. 2011;184:582–589. doi: 10.1164/rccm.201011-1909OC. - DOI - PubMed
    1. Weiden MD, et al. Biomarkers of World Trade Center Particulate Matter Exposure: Physiology of Distal Airway and Blood Biomarkers that Predict FEV(1) Decline. Semin Respir Crit Care Med. 2015;36:323–333. doi: 10.1055/s-0035-1547349. - DOI - PMC - PubMed
    1. Banauch GI, et al. Pulmonary function after exposure to the World Trade Center collapse in the New York City Fire Department. Am J Respir Crit Care Med. 2006;174:312–319. doi: 10.1164/rccm.200511-1736OC. - DOI - PMC - PubMed
    1. Prezant DJ, et al. Cough and bronchial responsiveness in firefighters at the World Trade Center site. N Engl J Med. 2002;347:806–815. doi: 10.1056/NEJMoa021300. - DOI - PubMed

Publication types

MeSH terms