Coffee is a worldwide beverage of increasing consumption owing to its unique flavor and several health benefits. The current research aimed at investigating multiple bioactive compounds in coffee belonging to the two major species of Coffea, including C. arabica and C. canephora or C. Robusta, either green or roasted, along with 15 commercial specimens collected from the Middle East. Two platforms were employed, including ultra-high-performance liquid chromatography coupled to mass spectrometry (UHPLC/MS) and UV for profiling and fingerprinting, respectively. Identification of the UPLC/MS dataset was aided by using the global natural product social molecular networking (GNPS). A total of 91 peaks were annotated belonging to different metabolites classes, including 23 hydroxycinnamates (i.e., chlorogenic acid), 5 hydroxycinnamic amides (i.e., Coumaroyl tryptophan), 17 diterpenes (i.e., Cafestol), 15 fatty acids (i.e., Trihydroxy-octadecanoic acid), 7 sphingolipids, 6 nitrogenous compounds (i.e., N-tricosanoyl-hydroxytryptamine), 2 alkaloids (i.e., Caffeine), 2 sugars (i.e., Acetyl-di-feruloyl sucrose),1 coumarin (dihydroxypsoralen-O-hexoside), and 1 fatty acid amides (docosenamide). Several metabolites are reported for the first time in coffee, including novel hydroxycinnamates (5), sphingolipids (4) and ceramides (4), diterpenes (2), coumarin (1), fatty acid amide (1).

Spectral datasets from both UV and UPLC/MS were subjected to multivariate data analysis for discrimination between the 19 coffee accessions, including principal component analysis (PCA) and orthogonal projection to latent structures-discriminate analysis (OPLS-DA). PCA model of the UV dataset (200-450 nm) provided effective discrimination of the unroasted from the roasted ones and further to classify closely related samples based on their origin. Moreover, UPLC/MS models could further distinguish between green and roasted and also roasted and instant samples. The results of both platforms were comparable, suggesting the UV technique as an alternative tool for UPLC/MS. Moreover, total phenolics assay (TPC) and antioxidant assays (DPPH and FRAP) were employed to correlate between both assays and UPLC/MS semi-quantitative dataset using partial least squares-discriminant (PLS), with a positive correlation between metabolites and assays. Potential markers that were revealed from the PLS model included caffeoylquinic acid, dihydroxypsoralen-O-hexoside, dicaffeoyl-quinolactone and caffeine. Moreover, mineral analysis was employed in different coffee specimens for major elements in coffee using inductively coupled plasma atomic emission spectrometry (ICP-AES). To the best of our knowledge, this is the first comprehensive comparative metabolomics approach targeting a large number of coffee specimens and to be used for future quality control determination in coffee revealing for seeds roasting indices or origin.


School of Sciences and Engineering


Chemistry Department

Degree Name

MS in Chemistry

Graduation Date

Summer 6-15-2021

Submission Date


First Advisor

Mohamed Farag

Committee Member 1

Tamer El-Idreesy

Committee Member 2

Sherweit H. El-Ahmady


124 p.

Document Type

Master's Thesis

Institutional Review Board (IRB) Approval

Not necessary for this item

Included in

Food Science Commons