Bioscience Journal | 2021 | vol. 37, e37072 | ISSN 1981-3163 1 Andre Dominghetti FERREIRA1 , Juliana Costa de Rezende ABRAHÃO2 , Gladyston Rodrigues CARVALHO2 , Alex Mendonça de CARVALHO3 , Vinicius Teixeira ANDRADE2 , Flavia Maria Avelar GONÇALVES4 , Marcelo Ribeiro MALTA2 1 Embrapa Café, Brazilian Agricultural Research Corporation, Brasilia, Distrito Federal, Brazil. 2 Epamig Sul, Agricultural Research Company of Minas Gerais, Lavras, Minas Gerais, Brazil. 3 Department of Agriculture, São Paulo State University, Registro, São Paulo, Brazil. 4 Department of Biology, Federal University of Lavras, Lavras, Minas Gerais, Brazil. Corresponding author: Andre Dominghetti Ferreira Email: andre.dominghetti@embrapa.br How to cite: FERREIRA, A.D., et al. Chemical and sensory characteristics in the selection of bourbon genotypes. Bioscience Journal. 2021, 37, e37072. https://doi.org/10.14393/BJ-v37n0a2021-54155 Abstract The evaluation of coffee quality in Brazil for commercialization is conducted mainly through sensory analysis, also known as the "cup test", in which professional tasters evaluate and score various attributes. The adoption of chemical methods could complement the sensory classification of beverages, if correlations between these chemical and sensory analyses exist, making classification less subjective. This work aimed to identify the relationships between the chemical and sensorial traits of coffee-beverage quality and to evaluate the use of these traits as criteria for the selection of Bourbon cultivars. Twenty coffee genotypes from the first three harvests across five municipalities of the state of Minas Gerais, Brazil were evaluated. The genotypic values, predicted for each genotype, were used to determine the index based on the sum of ranks from Mulamba and Mock. The genetic correlations among the evaluated traits were also estimated. The presented evaluations were not able to efficiently detect genetic and phenotypic relationships between the chemical and sensorial characteristics of drink quality, but as selection criteria for generation advancement in the beverage quality, it is possible to use these characteristics. Bourbon Amarelo LCJ 9-IAC, Bourbon Amarelo-Procafé, Bourbon Amarelo-Boa Vista, Bourbon Vermelho-São João Batista, and Bourbon Amarelo-Samambaia were the genotypes with the most promising cup quality in the studied regions. Through the selection of these five genotypes, the selection gain was 1.65% for sensory score for beverage quality, when the interaction among the studied environments was removed. The heritability was 92% for improving this trait. Keywords: BLUP. Coffea arabica. Genetic correlation. Genetic improvement. Simultaneous selection. 1. Introduction Given the change in the profile of coffee consumers associated with greater demand in the specialty coffee market, it is necessary to identify genotypes that present aptitude for the production of differentiated coffees. For this reason, researchers have been emphases beverage quality in the development of new coffee cultivars (Figueiredo et al. 2013; Sunarharum et al. 2014). Arabica coffee is known for producing brews with high cup quality. However, the chemical composition, is affected by a range of parameters, including cultivar, planting origin, harvest method, type CHEMICAL AND SENSORY CHARACTERISTICS IN THE SELECTION OF BOURBON GENOTYPES https://orcid.org/0000-0002-6512-4923 https://orcid.org/0000-0002-6026-7634 https://orcid.org/0000-0002-4466-674X https://orcid.org/0000-0002-2720-1318 https://orcid.org/0000-0002-0011-2788 https://orcid.org/0000-0003-4652-1824 https://orcid.org/0000-0003-0683-1078 Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 2 Chemical and sensory characteristics in the selection of bourbon genotypes of preparation, grain appearance, among others, that may condition a pleasant sensory profile in the cup (Toledo et al. 2016; Fassio et al. 2020). The Bourbon is a variety of the species Coffea arabica, which is known for its sensory characteristics making it prized and desirable in the specialty coffee industry (Figueiredo et al. 2013). This group of varieties produces high-quality beans for use in specialty coffees in diverse regions of the world (Flambeau et al. 2017; Steen et al. 2017). The sensory evaluation for commercialization is conducted mainly through physical examination and sensory analysis, also known as the "cup test", in which professional tasters evaluate and score various attributes (Pimenta et al. 2018). However, this process can lead to distortions, causing disagreement between samples tested by different tasters. The adoption of chemical analyses could complement the sensory classification of beverage, making classification less subjective, if correlations between the sensory and chemical trait analyses exist. There are few studies to date that correlate the physicochemical and sensory attributes of coffee and that use multivariate tools to understand coffee flavor (Farah et al. 2006; Alves et al. 2018). The selection of a set of important characters is necessary, aiming at adequate gains, simultaneously, in all characteristics in the breeding programs (Ashikaga et al. 2016). Selection indices are multivariate techniques that associate information related to various traits of agronomic interest with the genetic properties of the evaluated population. The index by Mulamba and Mock (1978) based on the sum of the rankings provides a simultaneous selection of gains in several situations. Furthermore, the mixed model method (Henderson 1984) is a procedure that allows obtaining estimates of genetic values and parameters and maximizes genetic gains with selection (Salgado et al. 2014; Arendacká and Puntanen 2015). Studies involving these procedures can contribute significantly to genetic improvement and, consequently, reduce the time spent in obtaining cultivars with better cup quality. The aims in this work were i) to identify the relationships between the chemical and sensorial traits of coffee-beverage quality and ii) to evaluate the use of chemical and sensorial traits as a criterion for selection of Bourbon coffee variety. 2. Material and Methods The experiments were conducted in two important coffee regions of the state of Minas Gerais (southern Minas Gerais and Alto Paranaíba) to represent the environmental conditions existing in regions known to produce fine coffees. The municipalities of Campos Altos, Santo Antônio do Amparo, Patrocínio, Lavras and, Três Pontas were selected (Table 1). The soils in all the selected regions were classified as Dystroferric Red Latosol. Table 1. Geographic region, climatic variables and, characterization of the experimental installation sites in the state of Minas Gerais. Municipalities Locations Altitude Temperature* Precipitation* Coordinates Campos Altos Alto Paranaíba 1,230 m 17,6°C 1,830 mm 19°41’46”S 46°59’33”N Santo Antônio do Amparo South of Minas 1,050 m 19,8°C 1,670 mm 20°56’47”S 44°55’08”O Patrocínio Alto Paranaíba 966 m 22°C 1,620 mm 18°56’38”S 46°59’33”N Lavras South of Minas 950 m 19,3°C 1,529 mm 21°14’43”S 44°59’59”O Três Pontas South of Minas 905 m 18°C 1,545 mm 21°20’50”S 45°28’23”O *Annual averages. There were evaluated twenty coffee genotypes obtained from different farms that had their products well rated in drinking quality contests. These farms are located in traditional coffee growing regions but have no records of the origin of cultivars that were originally planted. Therefore, the cultivars were denominated by the name of the farm: Bourbon Amarelo LCJ 10-Epamig de Machado (BA1), Bourbon Amarelo-Procafé Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 3 FERREIRA, A.D., et al. (BA2), Bourbon Amarelo-Bom Jardim (BA3), Bourbon Amarelo-Betânia (BA4), Bourbon Amarelo-Boa Vista (BA5), Bourbon Amarelo LCJ 9-IAC (BA6), Bourbon Amarelo-Toriba (BA7), Bourbon Amarelo-São Paulo (BA8), Bourbon Amarelo-Castro (BA9), Bourbon Amarelo-Nogueira (BA10), Bourbon Amarelo-Paixão (BA11), Bourbon Amarelo-Samambaia (BA12), Bourbon Vermelho-Procafé (BV13), Bourbon Vermelho-São João Batista (BV14), Bourbon Amarelo Italiano-Monte Alegre (BA15), Bourbon Amarelo Trigo- Monte Alegre (BA16) and Bourbon Amarelo Limoeiro-Monte Alegre (BA17). The cultivars Mundo Novo IAC 502/9 (S18), Catuaí Vermelho IAC 144 (S19) and Icatu Amarelo IAC 3282 (S20), which are widely cultivated throughout the state of Minas Gerais, were used as standards in the trials. A randomized complete block design was used, which included three replicates and plots consisting of 10 plants, of which the six central plants were used. The adopted spacing was 3.5 m between rows and 0.70 m between plants (4,081 plants ha-1). The analyses of chemical and sensorial attributes were performed annually during the first three harvests. After careful harvesting, the fruits in the cherry stage and the dried fruits were separated based on density of fruits in a water box and a sieve made of 3.0 mm x 3.0 mm wire mesh. The separation of fruits in the cherry stage from those in the green stage, which may have remained in the sample, was performed with the aid of a coffee peeler, which, by means of pressure exerted on the fruits, allowed only the ripe fruit to be peeled. Using this method, seven liters of peeled cherry coffee was obtained. These samples were uniformly distributed in 1 m2 sieves (with a wooden frame and a mesh screen of 2.00 x 1.00 mm, made of polyethylene wires), where they were rotated 12 times per day until the coffee beans reached a moisture content of approximately 11% (wb). After drying, the samples were organized and prepared for the chemical and sensorial analyses. The processed samples were frozen in liquid nitrogen and ground using IKA A11 Basic Analytic® mill, sprayed and kept at -80 °C for further analysis of total titratable acidity (ATT), reducing sugars (AR), nonreducing agents (ANR), potassium ion leaching (LK), electrical conductivity (CE), soluble solids (SS), total chlorogenic acids (ACT) and total phenolic compounds (CPT). These analyses were performed in triplicate and the data were expressed in g 100 g-1 (d.b.). The CE and the amount of LK were determined according to a methodology proposed by Prete and Abrahão (1995), with a samples soaking time of five hours. The acidity was determined by titration with 0.1 N NaOH according to a technique described by the AOAC (1990) and expressed in ml of 0.1 N NaOH per 100 grams sample. The sugars were extracted by the Lane-Enyon method, as described by the AOAC (1990), and determined by the Somogy technique as modified by Nelson (1944). To determine the SS, the grains were crushed, water was added and then the mixture was filtered as described by AOAC (1990). The amount of ACT was determined according to Abrahão et al. (2008). The CPT were extracted by the method of Goldstein and Swain (1963) and quantified by the method of Folin Denis (AOAC 1970). For sensory analysis, a panel of three trained judges (Q-grader) evaluated five cups of each sample in relation to eight attributes, clean cup (Ccp), sweetness (Swt), acidity (Acd), body (Bod), flavor (Fla), aftertaste (Aft), balance (Bln), and overall impression (Ove), and each attribute was assigned a score ranging from 0–8 based on the intensity of the sample, thus ensuring greater objectivity than conventional "cup tests". The sum of the attribute scores represented the final score of beverages (Fsc), allowing for the final classification of the drink. Each sample began with a pre-established score of 36 points, which incorporated the notes of each attribute, and samples with a score higher than 80 were classified as specialty coffee (Carvalho et al. 2016). The averages of quality attributes of first three harvests of each crop were used in the statistical analyses. The data obtained were adjusted to the following linear mixed model: y = Xr + Zg + Wb + e, which y is the data vector; r is the fixed-effects vector of the measurement-replication combinations added to the general mean; g is the vector of the genotypic effects N (0, A𝜎𝑔 2σ), which A is the kinship matrix and 𝜎𝑔 2σ is the genotypic variance; b is the vector of the effects of blocks N (0, I𝜎𝑏 2σ), which I is the identity matrix and 𝜎𝑏 2σ is the environmental variance between blocks; e is the vector of errors or residues N (0, I𝜎𝑒 2σ), where I𝜎𝑒 2σ is the residual variance; and X, Z, W are the incidence matrices for the effects r, g and b, respectively. Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 4 Chemical and sensory characteristics in the selection of bourbon genotypes Estimates of the genetic parameters were obtained by the REML/BLUP (restricted maximum likelihood/best linear unbiased prediction) procedure using the SELEGEN-REML/BLUP computational software (Resende 2016). The evaluation of the variances associated with the random effects were made using the likelihood ratio test at p<0.01 and p<0.05. The genotypic values, predicted for each genotype, were used to determine the index based on the sum of ranks from Mulamba and Mock (1978). The genetic correlations among the evaluated traits were also estimated using the Genes program (Cruz 2013). 3. Results The correlations between AR and ATT, CE and ATT, AR and Fsc as well as Fsc and CE were negative (Figure 1). These correlations were significant, but coefficients varying between 0.35 and 0.48. Therefore, they are not mentioned in the discussion. The characteristics AR and CE, RA and CPT, Fsc and ATT, and LK and CPT presented positive correlations of moderate to high magnitude, between 0.60 and 0.89. Figure 1. Pearson correlation matrix of the studied variables, including titratable total acidity (ATT), reducing sugars (AR), nonreducing sugars (ANR), potassium ion leaching (LK), electrical conductivity (CE), soluble solids (SS), total chlorogenic acids (ACT), total phenolic compounds (CPT) and final drink score (Fsc). The genotype variance indicates the possibility of selection for the ATT, ANR, LK, and Fsc characteristics (Table 2). The heritability estimates obtained for Fsc indicate that this trait can be selected for with high accuracy of 96% and reflected the quantity and quality of information and the procedures used in the prediction of genetic values. The ATT, ANR and LK characteristics presented moderate accuracy magnitudes varying from 49% to 66%. The genotypic correlation in environments was 83 and 99%, respectively, for the characteristics LK and Fsc indicating the agreement of the best progenies in the five evaluation municipalities. The observed values of the coefficient of determination due to the effects of the genotype x local interaction were 0.09 for the ATT and 0.29 for SS characters, values compatible with experiments considered accurate by Freitas et al. (2007). Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 5 FERREIRA, A.D., et al. Table 2. Estimates of the genotypic parameters related to titratable total acidity (ATT), reducing sugars (AR), nonreducing sugars (ANR), potassium ion leaching (LK), electrical conductivity (CE), soluble solids (SS), total chlorogenic acids (ACT), total phenolic compounds (CPT) and final score of beverage (Fsc). Parameters ATT AR ANR LK CE SS ACT CFT Fsc g v 4.73* 0.05 15.84 * 4.09* 0.01 6.33 0.01 6.03 4.20* 2 in t c 0.29 0.01 0.03 0.01 0.01 0.09 0.01 0.03 0.00 gen Ac 0.66 0.11 0.49 0.53 0.10 0.30 0.11 0.31 0.96 2 mg h 0,43 0,01 0,24 0,28 0,01 0,09 0,01 0,10 0,92 lo c rg 0.21 0.07 0.40 0.83 0.04 0.08 0.09 0.19 0.99 Average 205.37 5.75 75.74 26.91 72.36 46.31 7.91 34.86 83.39 *Significant at 5% as determined by the likelihood ratio test. g v Genotype variance, 2 in t c coefficient of determination for the effects of genotype x local interaction, 2 mg h heritability of genotype average, assuming complete survival, gen Ac accuracy of genotype selection, assuming complete survival, lo c rg genotype correlation between performance in environments. The choice of ATT, ANR, LK, and Fsc for selection of progenies was based on genotypic parameters. The progenies BA6, BA2, BA5, BA14, and BA12 were more promising for this characteristic (Table 3). The selection of these genotypes was based on the ability of characteristics to be simultaneously selected as described by Mulamba and Mock (1978). None of cultivars used as standards were among the genotypes selected (Table 3). The predicted genotypic values indicated improvement over the genotypic mean of each characteristic, with a 1.65% increase in Fsc, 0.38% increase in ATT, 1.26% increase in ANR, and 1.11% reduction in LK. Table 3. Estimates of average components using the BLUP procedure. The predicted additive genetic value (û + â) values related to the Mulamba and Mock (1978) index (Ij) related to titratable total acidity (ATT), nonreducing sugars (ANR), potassium ion leaching (LK), and final score of beverage (Fsc). Parameters ATT ANR LK Fsc Genotypes û + â û + â û + â û + â Ij BA1 205.79 74.94 74.94 85.66 24.53 BA2 205.79 75.58 75.58 84.73 26.18 BA3 205.21 73.15 73.15 85.66 28.32 BA4 206.36 74.38 74.38 84.59 25.72 BA5 206.51 75.66 75.66 85.27 26.88 BA6 206.51 78.24 78.24 86.61 25.95 BA7 204.06 74.59 74.59 83.89 27.64 BA8 205.38 77.41 77.41 84.81 27.30 BA9 206.22 76.64 76.64 84.24 26.12 BA10 206.65 72.16 72.16 82.08 27.77 BA11 205.50 73.34 73.34 81.41 26.54 BA12 207.52 77.39 77.39 84.32 26.27 BA13 206.22 78.28 78.28 84.75 28.80 BA14 204.49 76.67 76.67 82.94 27.77 BA15 203.77 76.00 76.00 80.59 26.43 BA16 203.19 74.30 74.30 80.40 28.04 BA17 203.05 78.72 78.72 80.52 25.70 S181 202.33 73.09 73.09 80.77 27.36 S191 206.94 77.10 77.10 82.30 26.79 S201 205.93 77.25 77.25 82.30 28.15 Average 205.37 75.74 26.91 83.39 GS%2 0.38 1.26 -1.11 1.65 1Cultivars used as standards. Overall average of the experiment. 2Predicted additive genetic gain considering the genotypes BA6, BA2, BA5, BA14, and BA12. Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 6 Chemical and sensory characteristics in the selection of bourbon genotypes 4. Discussion In order to know the associations of inheritable nature to be used in the orientation of breeding programs aiming at coffee-beverage quality, the genotypic correlation between the characteristics studied was estimated (Figure 1). ATT had a favorable effect on Fsc, indicating that acids not only contribute to sourness of coffee-beverage quality, but also appear to be important factors in determining particular flavor, and aroma of beverage (Clemente et al. 2015). AR was highly correlated with content of CE. This occurs, due the membrane degeneration, caused by possible fermentation due to sugars present in the bark, resulted in aggregation of leached electrolytes and coffee-beverage quality (Caixeta et al. 2013). The CPT showed a correlation with CE, and also with AR, which corroborates the findings of other authors (Chagas and Malta 2008). Phenolic components refer to flavonoids, phenolic acids, and phenolic diterpenes and have antioxidative effects that retard oxidative degradation of lipids and thereby improve the quality and nutritional value of food (Javanmardi et al. 2003). The correlations observed in this work suggest that the chemical analyses of total titratable acidity and electrical conductivity do not aid in the identification of genotypes that result in a specialty beverage. Instead, these correlations demonstrate that titratable acidity and electrical conductivity activities are related to processes that respond to grain damage (Taveira et al. 2015), resulting in a measurable loss of quality that was not representative of a change in the appearance of grains. Despite the wide application of chemometrics, correlating compositional data with sensory attributes is a complicated task (Sunarharum et al. 2014). Under optimal harvesting and post harvesting processes, as in the present work, these chemical evaluations may not be able to detect relationships between the chemical and sensorial quality characteristics of the beverage. The implemented analysis allowed the simultaneous estimation of genotype mean heritability and genotype correlation between performance in environments. It also allowed identifying progenies coffee- beverage quality confirmed genetic gain, and, thus, supported our hypothesis of the use of chemical and sensorial traits as a selection criterion. The genetic gain is inversely proportional to intensity of selection, which determines the number of selected individuals. Thus, in the present study, the need to work with a large number of individuals (five genotypes, 25% selection intensity) to ensure a minimum effective number that, according to Rocha et al. (2009), allows greater efficiency in the subsequent selection stages was considered (Table 3). When aiming for the beverage quality ideotype, superior genotypes for all characteristics studied are sought. The simultaneous selection criteria applied to the genetic values favored positive genetic gains in the collection of genotypes Bourbon Amarelo LCJ 9-IAC, Bourbon Amarelo-Procafé, Bourbon Amarelo-Boa Vista, Bourbon Vermelho-São João Batista and Bourbon Amarelo-Samambaia for indicating the potential use of these cultivars for the advancement of generations. Similarly, Ferreira et al. (2012), analyzed the sensory beverage quality of these genotypes and determined that the genotype Bourbon Amarelo LCJ 9-IAC had the greatest potential for the production of specialty coffees in the five environments studied. The selected five genotypes had scores between 82 and 86 (Table 3). In order for a coffee to be classified as a specialty coffee, its score should be higher than 80 (Chalfoun et al. 2013). Their genetic potential to produce specialty coffees is clear, as determined by predicted additive genetic value. These genotypes can be used for production of coffee targeting a different market from commodity coffee and meeting the demand of specialty coffee growers and buyers. In this study, we proposed ways to improve coffee-beverage quality plant selection. Five genotypes with quality aptitude showing selection gain of 1.65% for sensory score for beverage quality were selected. It is also worth noting the possibility of improving the final drink through the selection of these five genotypes; when the interaction among the studied environments was removed, there was 92% heritability for improving sensory score for beverage quality (Table 2). We reinforce the superiority of BA14 genotype which was previously evaluated in field conditions by Ferreira et al. (2013). This genotype has desirable agronomic characteristics, and also stability and adaptability in different environments. The superior genotypes identified are promising options for integration into coffee breeding programs focused on quality and should be selected by the breeder considering the overall need for desired traits. Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 7 FERREIRA, A.D., et al. 5. Conclusions The presented evaluations were not able to efficiently detect genetic and phenotypic relationships between the chemical and sensorial characteristics of drink quality, but as selection criteria for generation advancement in the beverage quality, it is possible to use these characteristics. Bourbon Amarelo LCJ 9-IAC, Bourbon Amarelo-Procafé, Bourbon Amarelo-Boa Vista, Bourbon Vermelho-São João Batista, and Bourbon Amarelo-Samambaia were the genotypes with the most promising cup quality in the studied regions. Through the selection of these five genotypes, the selection gain was 1.65% for sensory score for beverage quality, when the interaction among the studied environments was removed. The heritability was 92% for improving this trait. Authors' Contributions: FERREIRA, A.D.: conception and design, acquisition of data, analysis and interpretation of data, critical review of important intellectual content; ABRAHÃO, J.C.R.: conception and design, acquisition of data, analysis and interpretation of data, drafting the article; CARVALHO, G.R.: conception and design, acquisition of data, analysis and interpretation of data, critical review of important intellectual content; CARVALHO, A.M.: conception and design, acquisition of data, analysis and interpretation of data, critical review of important intellectual content; ANDRADE, V.T.: conception and design, acquisition of data, analysis and interpretation of data, critical review of important intellectual content; GONÇALVES, F.M.A.: conception and design, acquisition of data, analysis and interpretation of data, critical review of important intellectual content; MALTA, M.R.: conception and design, acquisition of data, analysis and interpretation of data, critical review of important intellectual content. All authors have read and approved the final version of the manuscript. Conflicts of Interest: The authors declare no conflicts of interest. Ethics Approval: Not applicable. Acknowledgments: The authors would like to thank the funding for the realization of this study provided by the Brazilian agencies FAPEMIG (Fundação de Amparo à Pesquisa do Estado de Minas Gerais - Brasil), CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico - Brasil), and the Coffee Research Consortium-Consórcio Pesquisa Café, National Institute of Coffee Science and Technology-INCT Café/CNPq. References ABRAHAO, S.A., et al. Bioactive compounds in whole and decaffeinated coffee and sensorial quality of the beverage Pesquisa Agropecuária Brasileira. 2008, 43(12), 1799-1804. http://dx.doi.org/10.1590/S0100-204X2008001200022 ALVES, A., et al. Influence of environmental and microclimate factors on the coffee beans quality (Coffee canephora): Correlation between chemical analysis and stable free radicals. Agricultural Sciences. 2018, 9(9), 1173-1187. http://dx.doi.org/10.4236/as.2018.99082 ARENDACKÁ, B. and PUNTANEN, S. Further remarks on the connection between fixed linear model and mixed linear model. Statistical Papers. 2015, 56(4), 1235–1247. https://doi.org/10.1007/s00362-014-0634-2 ASHIKAGA, K., et al. Simultaneous selection for nutritive value and agronomic traits in timothy (Phleum pratense L.). Euphytica. 2016, 208(2), 237-250. https://doi.org/10.1007/s10681-015-1583-0 AOAC. Official methods of analysis of the Association of Official Analytical. 15th ed. Washington, DC: Association of Official Analytical Chemistry, 1990. CAIXETA, I.F., GUIMARÃES, R.M. and MALTA, M.R. Quality of coffee seeds after retardment of post-harvest processing. Coffee Science. 2013, 8(3), 249-255. https://doi.org/10.25186/cs.v8i3.425 CARVALHO, A.M. de, et al. Relationship between the sensory attributes and the quality of coffee in different environments. African Journal of Agricultural Research. 2016, 11(38), 3607-3614. https://doi.org/10.5897/AJAR2016.11545 CHAGAS, S.J.R. and MALTA, M.R. Avaliação da composição química do café submetido a diferentes formas de preparo e tipos de terreiros de secagem. Revista Brasileira de Armazenamento. 2008, 10(2), 1-8. CHALFOUN, S.M., et al. Sensorial characteristics of coffee (Coffea arabica L.) varieties in the Alto Paranaíba. Coffee Science. 2013, 8(1), 43-52. https://doi.org/10.25186/cs.v8i1.330 CLEMENTE, J.M., et al. Effects of nitrogen and potassium on the chemical composition of coffee beans and on beverage quality. Acta Scientiarum: Agronomy. 2015, 37(3), 297-305. https://doi.org/10.1590/S0100-204X2018000400006 CRUZ, C.D. Genes: a software package for analysis in experimental statistics and quantitative genetics. Acta Scientiarum: Agronomy. 2013, 35(3), 271-276. http://dx.doi.org/10.4025/actasciagron.v35i3.21251 FASSIO, L. de O., et al. Performance of arabica coffee accessions from the active germplasm bank of Minas Gerais – Brazil as a function of dry and wet processing: a sensory approach. Australian Journal of Crop Science. 2020, 14(6), 1011-1018. https://doi.org/10.21475/ajcs.20.14.06.p2528 FARAH, A., et al. Correlation between cup quality and chemical atributes of Brazilian coffee. Food Chemistry. 2006, 98(2), 373-380. https://doi.org/10.1016/j.foodchem.2005.07.032. http://dx.doi.org/10.1590/S0100-204X2008001200022 http://dx.doi.org/10.4236/as.2018.99082 https://doi.org/10.1007/s00362-014-0634-2 https://doi.org/10.1007/s10681-015-1583-0 https://doi.org/10.25186/cs.v8i3.425 https://doi.org/10.5897/AJAR2016.11545 https://doi.org/10.25186/cs.v8i1.330 https://doi.org/10.1590/S0100-204X2018000400006 http://dx.doi.org/10.4025/actasciagron.v35i3.21251 https://doi.org/10.21475/ajcs.20.14.06.p2528 Bioscience Journal | 2021 | vol. 37, e37072 | https://doi.org/10.14393/BJ-v37n0a2021-54155 8 Chemical and sensory characteristics in the selection of bourbon genotypes FERREIRA, A.D., et al. Análise sensorial de diferentes genótipos de cafeeiros Bourbon. Interciencia. 2012, 37(5), 390-394. FERREIRA, A.D., et al. Desempenho agronômico de seleções de café Bourbon Vermelho e Bourbon Amarelo de diferentes origens. Pesquisa de Agropecuária Brasileira. 2013, 48(4), 388-394. https://doi.org/10.1590/S0100-204X2013000400006 FIGUEIREDO, L.P., et al. The Potential for High Quality Bourbon Coffees From Different Environments. Journal of Agricultural Science. 2013, 5(10), 87-98. https://doi.org/10.5539/jas.v5n10p87 FLAMBEAU, K.J., LEE W.J. and YOON J. Discrimination and geographical origin prediction of washed specialty Bourbon coffee from different coffee growing areas in Rwanda by using electronic nose and electronic tongue. Food Science Biotechnology. 2017, 26(5), 1245-1254. https://doi.org/10.1007/s10068-017-0168-1 FREITAS, Z.M.T.S. de, et al. Evaluation of quantitative traits related with the vegetative growth among arabica coffee cultiv ars. Bragantia. 2007, 66(2), 267-275. https://doi.org/10.1590/S0006-87052007000200010 GOLDSTEIN, J.L. and SWAIN, T. Changes in tannins in ripening fruits. Phytochemistry. 1963, 2(4), 371-383. https://doi.org/10.1016/S0031- 9422(00)84860-8 HENDERSON, C.R. Applications of linear models in animal breeding. 1st ed. Ontario: Guelph, 1984. JAVANMARDI, J., et al. Antioxidant activity and total phenolic content of Iranian Ocimum accessions. Food Chemistry. 2003, 83(4), 547-550. http://dx.doi.org/10.1016/S0308-8146(03)00151-1 MULAMBA, N.N. and MOCK J.J. Improvement of yield potential of the Eto Blanco maize (Zea mays L.) population by breeding for plant traits. Egypt Journal of Genetics and Cytology. 1978, 7(1), 40-51. NELSON, N. A Photometric adaptation of Somogy method for the determination of glucose. Journal of Biological Chemists. 1944, 153(1), 375- 384. https://doi.org/10.1016/S0021-9258(18)71980-7 PIMENTA, C.J., ANGÉLICO, C.L. and CHALFOUN, S.M. Challenges in coffee quality: Cultural, chemical and microbiological aspects. Ciência e Agrotecnologia. 2018, 42(4), 337-349. http://dx.doi.org/10.1590/1413-70542018424000118 PRETE, C.E.C. and ABRAHÃO, J.T.M. Condutividade elétrica dos exsudatos de grãos de café (Coffea arabica L.) I Desenvolvimento da Metodologia. Semina. 1995, 16(1), 17-21. RESENDE, M.D.V. Software Selegen-REML/BLUP: a useful tool for plant breeding. Crop Breeding and Applied Biotechnology. 2016, 16(4), 330- 339. https://doi.org/10.1590/1984-70332016v16n4a49 ROCHA, R.B., et al. Avaliação genética de procedências de bandarra (Schizolobium amazonicum) utilizando REML/BLUP (Máxima verossimilhança restrita/Melhor predição linear não viciada). Scientia Forestalis. 2009, 37(84), 351-358. SALGADO, S.M.L., REZENDE, J.C. de and NUNES, J.A.R. Selection of coffee progenies for resistance to nematode Meloidogyne paranaensis in infested area. Crop Breeding and Applied Biotechnology. 2014, 14(2), 94-101. http://dx.doi.org/10.1590/1984-70332014v14n2a17 STEEN, I., et al. Influence of serving temperature on flavour perception and release of Bourbon Caturra coffee. Food Chemistry. 2017, 219, 61- 68. http://dx.doi.org/10.1016/j.foodchem.2016.09.113 SUNARHARUM, W.B., WILLIAMS, D.J. and SMYTH, H.E. Complexity of coffee flavor: A compositional and sensory perspective. Food Research International. 2014, 62(1), 315-325. https://doi.org/10.1016/j.foodres.2014.02.030 TAVEIRA, H. da S.J, et al. Post-harvest effects on beverage quality and physiological performance of coffee beans. African Journal of Agricultural Research. 2015, 10(12), 1457-1466. https://doi.org/10.5897/AJAR2014.9263 TOLEDO, P.R., PEZZA, L., PEZZA, H.R. and TOCI, A.T. Relationship between the different aspects related to coffee quality and their volatile compounds. Comprehensive Reviews in Food Science and Food Safety. 2016, 15(4), 705-719. https://doi.org/10.1111/1541-4337.12205 Received: 25 April 2020 | Accepted: 30 August 2020 | Published: 29 December 2021 This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. https://doi.org/10.1590/S0100-204X2013000400006 https://doi.org/10.5539/jas.v5n10p87 https://doi.org/10.1007/s10068-017-0168-1 https://doi.org/10.1590/S0006-87052007000200010 https://doi.org/10.1016/S0031-9422(00)84860-8 https://doi.org/10.1016/S0031-9422(00)84860-8 http://dx.doi.org/10.1016/S0308-8146(03)00151-1 https://doi.org/10.1016/S0021-9258(18)71980-7 http://dx.doi.org/10.1590/1413-70542018424000118 https://doi.org/10.1590/1984-70332016v16n4a49 http://dx.doi.org/10.1590/1984-70332014v14n2a17 http://dx.doi.org/10.1016/j.foodchem.2016.09.113 https://doi.org/10.1016/j.foodres.2014.02.030 https://doi.org/10.5897/AJAR2014.9263 https://doi.org/10.1111/1541-4337.12205