Cassia A. Tavares's research while affiliated with University of Campinas and other places

The ripe Brazilian tomato cultivar 'Santa Cruz' (10 sample lots) presented a mean composition in μg/g of 3.7 cis-phytofluene, 5.1 trans-β-carotene, 0.4 trans-ζ-carotene, 0.7 trans-γ-carotene, 3.0 cis-lycopene and 31.1 trans-lycopene. Cis-isomers of β-carotene, not detected in the fresh tomato, appeared in the processed products (39 samples). The juice which received the mildest heat treatment had only 0.01-0.02 μg/g 13-cis-β-carotene. In the puree and paste, O.1 to 3.8 μg/g 13-cis-β-carotene and not detected to 1.1 μg/g 9-cis-β-carotene were encountered. In ketchup, the ranges were 0.2-0.6 μg/g 13-cis-β-carotene and 0.4-0.8 μg/g 9-cis-βcarotene. The vitamin A value ranges (RE/100 g) were 63-101, 25-38, 46-160, 54-203 and 54-79 for fresh tomato, juice, puree, paste and ketchup, respectively. The corresponding lycopene ranges (μg/g) were 12-73, 53-68, 56-273, 110-240 and 59-143. The tomato paste, the most concentrated product, had higher vitamin A value and lycopene content, but at lower levels than expected, indicating losses during the more drastic processing.

An attempt was made to estimate the error in the determination of the vitamin A value (activity) when the cis- and trans-isomers are not separated and quantified individually. In 10 sample lots of fresh tomatoes, cis-isomers were not detected. However, in 52 samples of tomato products, varying amounts of cis-isomers of β-carotene were found. Utilising the currently employed calculation procedure, overestimations of 8–13% in juice, 10–66% in puree and paste, 15–26% in ketchup were observed when isomer separation was not undertaken. When column loss was taken into consideration, these errors decreased to 4–8%, 6–62% and 11–21%, respectively. Using the biopotencies obtained by Sweeney and Marsh (1970), aside from column loss, the overestimations were 4–8%, 3–52% and 7–15%, respectively. Thus, the error can be considerable, but the magnitude depends on how the calculation is carried out, demonstrating the necessity of establishing the true biopotencies, aside from obtaining more efficient chromatographic separation and accurate quantitation.