Brucella abortis infection detection by GC/MS in bison

Blog Post created by r_s_stahl on Jan 22, 2016

I have recently had the opportunity to develop a GC/MS method to measure volatile organic compounds (VOC's) in breath samples collected from bison (Bison bison) as a means of discriminating Brucella-infected bison from uninfected bison. Brucella abortis is a bacterium that causes bovine brucellosis. This disease can cause abortions, infertility and lowered milk production in bison and cattle. Elk and bison populations in the Greater Yellowstone Area serve as reservoirs for Brucella abortis and may transmit the disease to cattle. Current detection methods for this disease rely on a suite of serology-based assays. The development of a non-invasive method using breath samples would be less stressful to the animals and potentially more cost effective.


  • Breath samples were collected from Brucella-infected, naturally exposed bison and uninfected bison from the same herd, using a Tenax cartridge by drawing an air sample from a reservoir placed over each bison's muzzle.
  • Analytes were solvent extracted and separated on a DB-5ms (J&W Scientific, Agilent Technologies, Santa Clara, CA) 30 m x 250 mm column with a film thickness of 0.25 mm using an Agilent 6890 GC coupled with an Agilent 5973 MS using a thermal gradient from 30 oC to 300 oC ramped at 5 oC/min.
  • Peaks were identified using the NIST W8N08 mass spectrum library and peak areas were determined for: Hexanal (rt = 4.3 min), 2-Methyl-octane (rt = 9.4 min), 2-Ethyl-1-hexanol (rt = 10.3 min), 1-Phenyl-ethenone (rt  = 11.3 min), Nonanal (rt  = 12.7 min), Butanoic acid, 2-methylpropyl ester (rt = 19.2 min), Propanoic acid, 2-methyl-2,2-dimethyl-1-(2-hydroxy-1-methylethyl) propyl ester (rt = 19.6 min), and Propanoic acid, 2-methyl-3-hydroxy-2,4,4-trimethylpentyl ester (rt = 20.3 min).
  • A principle component analysis (PCA) was applied to these peak areas and the resulting scores were incorporated into linear discriminant analysis (LDA) classification models to predict brucellosis status. The sensitivity and specificity of the models using this approach resulted in sensitivity values ranging from 55.4 to 73.0 %, and specificity values of 72.0 to 91.1 %.


The results from this method are promising and indicate that concentration changes for VOC's in breath samples, determined by GC/MS analysis may provide a means of noninvasively identifying Brucella-infected bison.


Mention of a product does not constitute endorsement.



Image of the sampling reservoir placed over the muzzle of a restrained bison.