The
quality of East African coffee beans has been significantly
reduced by a flavor defect known as potato taste defect (PTD) due
to the presence of 2-isopropyl-3-methoxypyrazine (IPMP) and 2-isobutyl-3-methoxypyrazine
(IBMP). Therefore, the aims of this study were to determine the correlation
between these methoxypyrazines and the severity of odor attributed
to PTD and discover additional analytes that may be correlated with
PTD using Fisher ratio analysis, a supervised discovery-based data
analysis method. Specialty ground roasted coffees from East Africa
were classified as clean (i.e., no off-odor), mild, medium, or strong
PTD. For the samples examined, IPMP was found to discriminate between
non-defective and defective samples, while IBMP did not do so. Samples
affected by PTD exhibited a wide range of IPMP concentration (1.6–529.9
ng/g). Except for one sample, the IPMP concentration in defective
samples was greater than the average IPMP concentration in the non-defective
samples (2.0 ng/g). Also, an analysis of variance found that IPMP
concentrations were significantly different based on the severity
of odor attributed to PTD (p < 0.05). Fisher ratio
analysis discovered 21 additional analytes whose concentrations were
statistically different based on the severity of PTD odor (p < 0.05). Generally, analytes that were positively correlated
with odor severity generally had unpleasant sensory descriptions,
while analytes typically associated with desirable aromas were found
to be negatively correlated with odor severity. These findings not
only show that IPMP concentration can differentiate the severity of
PTD but also that changes in the volatile analyte profile of coffee
beans induced by PTD can contribute to odor severity.