<![CDATA[Tithonia diversifolia, known as Mexican sunflower, has been widely used as an herbal medicine to treat diabetes. This study used FTIR fingerprint spectra combined with chemometrics to differentiate T. diversifolia leaves extracts with different extracting solvents and their correlation with the inhibition of α-glucosidase activity. T. diversifolia collected from two growing locations (West Bandung and Sleman, Indonesia) was extracted with absolute ethanol, 50% ethanol, and water using ultrasonication. The ethanol absolute extract yielded a higher IC50 than the 50% ethanol and water extract. The FTIR spectra of each extract had a different profile, implying that the composition and the concentration of the metabolite extracted were relatively distinct. Absorbance data from the FTIR spectra in the 4000–400 cm−1 range were used to group all extracts according to the extracting solvent using principal component analysis (PCA). Before PCA, the FTIR spectra were subjected to signal preprocessing using a standard normal variate. We found that all of the extracts could be distinguished based on the extracting solvents using principal components (PC) 1 and 2 with a cumulative percentage of approximately 87%. Partial least square regression (PLSR) was used to correlate the FTIR spectra and the inhibition of the α-glucosidase activity to obtain a functional group of a metabolite that contributed to inhibiting the α-glucosidase activity. From the PLSR, peaks from the wavenumbers at ~3300 cm−1, ~3000 cm−1, ~1650 cm−1, ~1350 cm−1, and ~1100 cm−1 corresponded to the O-H, CH3, CH2, C=C, and C-O, which were thought to be responsible for inhibiting the α-glucosidase. Therefore, these functional groups were owned by the metabolites in the T. diversifolia leaves extracts that contributed to the inhibition of α-glucosidase.]]>
