05), except between dark and medium roasted filtered brews. Similar results were reported in a previous study by Tfouni et al. (2012) where no correlation was found between PAHs levels and the roasting degree of ground roasted coffee.

This is due to the high variability of the process, as shown by the results obtained within the same cultivar and roasting degree, submitted to the same brewing procedure. The coefficients of variation of the process replicates ranged from 12% (C. canephora cv. Apoatã, IDH tumor dark roasted, boiled) to 62% (C. canephora cv. Apoatã, medium roasted, filtered). This high variability is probably due to the roasting process since, although the temperature of the roaster was set at 200 °C, when green coffee beans are placed inside, the equipment suffers a temperature variation that is inherent to the roasting process. The internal temperature drops and then starts to increase again throughout the process. Although there was an effort to maintain the same roasting profile for replicates of all processes, some differences were observed, with some samples reaching higher temperatures

in a shorter/longer period of time than others ( Tfouni et al., 2012). Other authors presented results of PAHs levels in relation to coffee roasting process. Kayali-Sayadi, Rubio-Barroso, Cuesta-Jimenez, and Polo-Díez (1999) reported higher PAHs concentrations for brews made from commercial ground roasted coffees (2.87 ng/L) than the ones made from green or decaffeinated (1.99 and 1.65 ng/L, respectively),

Epigenetic inhibitor concentration there was no mention on the samples roasting degree. Houessou et al. (2007) did not detect or detected only traces of BbF, BkF and BaP in coffee brews prepared from ground coffees roasted for 5 min under different temperatures. BaA was detected in the range of traces to 0.15 μg/L (260 °C/5 min). The PAHs transfer to the coffee brew could be related to the known formation of a caffeine-PAHs complex (Kolarovic and Traitler, 1982, Moret and Conte, 2000 and Navarro et al., 2009). As C. canephora presents higher caffeine content than C. arabica, one should expect that the levels of PAHs in C. canephora brews would be higher due to the formation of the complex, which would facilitate the transfer of these lipophilic compounds to the brew. Nevertheless, in the Phospholipase D1 present study, coffee brews prepared with C. arabica cv. Catuaí Amarelo ground roasted beans presented mean summed PAHs levels higher than the ones prepared with C. canephora cv. Apoatã, independently of the brewing procedure used ( Fig. 1). C. arabica was contaminated with mean summed PAHs concentrations of 0.052 and 0.034 μg/L (filtered and boiled brews, respectively), while C. canephora presented 0.034 and 0.030 μg/L. This might be explained by the fact that the caffeine levels are much higher than the PAHs in both coffees (1195 mg/100 g, arabica; 1729 mg/100 g, canephora ( Tfouni et al.