A mulit-omics approach to finding potential biomarkers in Philippine civet coffee

Date of Publication


Document Type


Degree Name

Doctor of Philosophy in Chemistry

Subject Categories



College of Science



Thesis Adviser

Marissa G. Noel

Defense Panel Chair

Nancy Lazaro-Llanos

Defense Panel Member

Lourdes P. Guidote
Jaime Raul O. Janairo
Maria Carmen S. Tan
Rosario S. Sagum


Civet coffee, obtained from the feces of civet cats feeding on coffee cherries, is an exotic, high-priced beverage. But due to issues regarding counterfeiting, animal abuse, and disease spread, physico-chemical characterization and authentication studies have become necessary. So far the most extensive investigations prior to this study include scanning electron microscopy of the coffee beans, SDS-PAGE on green (raw unroasted) coffee beans electronic nose experiments on volatile compounds from the brew and a metabolomics approach that identified malic acid, citric acid, and the inositol/pyroglutamic acid ratio as possible discriminant markers for the authentication of civet coffee.

In this dissertation, physicochemical characterizations, microbial determination, and separation with mass spectrometric methods for the analyses of metabolites, endogenous oligosaccharides, proteins, and endogenous peptides were developed and carried out.

Civet coffee exhibits surface micro-pitting, as revealed by SEM analysis. This was illustrated by several dark regions representing cavities or pores and could be suggestive of reactions involving enzymes and / or acidic substances at the surfaces of the beans. These observations were consistent with an earlier SEM study on civet coffee. EDX analysis, meanwhile, showed that the mineral content of civet coffee was generally lower than that of the non - civet counterpart. This could be the result of absorption of major minerals in the digestive tract of the cat. Additionally, beans that passed through the civet’s body system were found to have higher caffeine content. It is known that caffeine is not only accumulated through the diet but it is known to be biosynthesized from purine nucleotides. The results of this study also show an observable decrease in the α-tocopherol content as the bean is processed inside the digestive system of the Civet cat which may have absorbed or partially digested the vitamin.

For aroma compound profiling, a total of 96 volatiles were determined. Almost all of the volatile compounds detected had been identified in previous studies.

Generally, no significant amount of microbes were found in roasted coffee. Much of this is attributed to the roasting process.

Using factorial analysis from metabolomics data, it can be inferred that among the three factors, treatment (i.e. roasting) has the single most significant effect on the metabolite profile of coffee and that, more notably, category (i.e. animal perturbation) has relatively little effect on the same.

From SDS-PAGE analysis of proteins, extracts from green beans exhibited notable decreases in band intensity, particularly for proteins between the 25 and 50 kDa region. The visual decrease was more pronounced in Robusta extracts (RNG vs RCG). These band intensity decreases corroborate earlier reports that proteolysis may have taken place while the bean was inside the animal’s gut.

In all green coffee beans the following peptides were present (liberica and robusta, civet and non-civet): AQYGAEYGNQK, AQYGAEYGNQKSQYDEYGNPVRQTD, ADLRDEYGNPMQLTDQYGNPVQLKDEYGNPMQLS, SLTGKLVNQIEIK, KAIILTTHSME. However, among the peptides in green coffee, perhaps the ones with greatest potential as biomarkers are the sequences whose manifestations in the triplicates are strong and consistent. These include: SLGGPSKLPFFDDPIALD, GVKSVEILEGDGGVT, GVKSVEILEGDGGVTIK, and GVKSVEILEGDGGVTIKLT.

Finally, a glycomics method was developed for coffee analysis. Tandem MS experiments were carried out on hot water-extracted coffee to determine the compositions of the glycans in all samples. As of this writing, there has been no report about oligosaccharide libraries or database for coffee. Initially, 328 compounds were detected from an initial filter using a lab-based unpublished software (Glyconote). Further individual spectral analysis trimmed down this list to 163, with a few still unconfirmed. Multivariate analysis is yet to be carried out to establish, if there are, differences in glycan profiles among samples. Monosaccharide and glycan quantitation demonstrated and corroborated changes in the glycan profiles of the samples.

Overall, results so far point to peptidomics as the most promising approach to finding biomarkers for the authentication of civet coffee.

Abstract Format






Accession Number


Shelf Location

Archives, The Learning Commons, 12F Henry Sy Sr. Hall


Coffee; Biochemical markers

This document is currently not available here.