Food Chemistry Advances 2 (2023) 100259

Contents lists available at ScienceDirect

Food Chemistry Advances

journal homepage: www.elsevier.com/locate/focha

Relationship between sensory and microbial profiles of fermented coffee by

dry and washed methods
Luiz Henrique Bozzi Pimenta de Sousa a, José Maria Rodrigues da Luz b,
Marliane de Cássia Soares da Silva c, Aldemar Polonini Moreli d, Tomás Gomes Reis Veloso c,
Rogério Carvalho Guarçoni e, Taís Rizzo Moreira f, Marcos Vinícius Pereira Barros c,
Maria Catarina Megumi Kasuya c, João Paulo Pereira Marcate d, Dério Brioschi Júnior d,
Willian dos Santos Gomes g, Lucas Louzada Pereira d,∗, Emanuele Catarina da Silva Oliveira d
a
Federal Institute of Espírito Santo (IFES), Coffee Design Group, Venda Nova do Imigrante, Espírito Santo. (Rua Elizabeth Minete Perim, S/N, Bairro São Rafael,
Espírito Santo-ES, 29375-000)
b
Universidade Federal de Viçosa (UFV), Departamento de Microbiologia, Laboratório de Associações Micorrizicas -LAMIC, Viçosa, Minas Gerais. (Avenida Ph Rolfs
S/N Viçosa, Minas Gerais-Mg 36570-000
c UFV - Departamento de Microbiologia, Laboratório de Associações Micorrizicas -LAMIC, Viçosa, Minas Gerais. (Avenida Ph Rolfs S/N Viçosa, Minas Gerais-MG

36570-000, Brazil
d
IFES, Coffee Design Group, Venda Nova do Imigrante, Espírito Santo. (Rua Elizabeth Minete Perim, S/N, Bairro São Rafael, Espírito Santo-ES, 29375-000), Brazil
e
Capixaba Institute of Technical Assistance, Research and Extension – INCAPER, Department of Statistics, Rua Afonso Sarlo, 160, Bento Ferreira, CEP: 29052-010,
Vitória, Espírito Santo, Brazil
f
Federal University of Espirito Santo (UFES), Department of Chemistry and Physic, Center for Exact, Natural and Health Sciences, Alto Universitário, sn, Guararema,
CEP: 29500-000, Alegre, Espírito Santo, Brazil
g
UFES, Department of Chemistry and Physic, Center for Exact, Natural and Health Sciences, Alto Universitário, sn, Guararema, CEP: 29500-000, Alegre, Espírito Santo,
Brazil

a r t i c l e i n f o a b s t r a c t

Keywords: Microorganisms influence the sensory quality of the coffee beverage through the compounds degradation or
Induced fermentation release of metabolites that diffuse into the fruit. However, type and duration of fermentation can contribute to
Chemical and sensory interactions the growth and diversity of microorganisms. Thus, the objective of this study was to analyze the microbial and
Flavor
sensory profile of Arabica coffees under different fruit-processing conditions and fermentation by Saccharomyces
Beverage
cerevisiae. Coffee cherries were used in the dry fermentation and washed fermentation at 28°C. Fermentation
DGGE
Q-Graders time ranged from 36 to 288 hours. A panel of sensory attributes was evaluated by 9 Q-graders. Mid-infrared
spectra were obtained in the wavenumber range from 4000 to 650 cm−1 . Microbial analyses were performed by
denaturing gradient gel electrophoresis. Total sensory score for fermented coffee was higher than for unfermented
coffee. In 108 and 216 hours of fermentation, the total score was higher in washed fermentation than in dry
fermentation, which also shows the influence of post-harvest processing and the removal of husk on the quality
of fermented coffee. Balance and body were the sensory attributes that most contributed to group formation in dry
fermentation. Thus, induced fermentation increases the sensory classification of coffee beverages with positive
effects on chemical and sensory profiles.

1. Introduction (Castellanos-Onorio et al., 2011, Pereira, Guarçoni, Moreira, Pimenta,

& De Sousa Cardoso, 2019, Pereira et al., 2020). Furthermore, the fi-
The post-harvest processing of coffee is a crucial step for the chemical nal quality of the coffee beverage may be determined by several fac-
composition of the bean and the sensory quality of the coffee beverage tors, such as climatic conditions (Bertrand et al., 2012), topographic
(da Silva Oliveira et al., 2022, da Silva et al., 2022). Over centuries of conditions (Silva, Queiroz, Ferreira, Corrêa, Rufino, 2016), genotypes
production, several techniques and protocols for the post-harvest pro- (Machado et al., 2021), stages of pre-harvest (Duguma & Meer, 2018),
cessing of coffee have been developed, disseminated, revised, and im- harvest management, post-harvest processing (Evangelista, Ferreira,
proved to achieve the chemical, nutritional, and sensory quality of coffee Marques, & Freitas, 2015), and roasting profile (Bhumiratana, Koushik,

∗
Corresponding author.
E-mail address: lucaslozada@hotmail.com (L.L. Pereira).

https://doi.org/10.1016/j.focha.2023.100259
Received 17 September 2022; Received in revised form 10 March 2023; Accepted 30 March 2023
2772-753X/© 2023 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

& Chambers, 2011), which culminate in different notes in the sensory sory quality because of the available resources and food safety (da Silva
analysis of the beverage (Pereira et al., 2017, Pereira et al., 2019). et al., 2022; De Bruyn et al., 2017).
In the post-harvest processing stages of coffee fruits, the methodolo- Thus, the objective of this study was to analyze the microbial and
gies used are dry (Dry fermentation) and wet (Washed fermentation) sensory profile of Arabica coffees under different fruit processing con-
processing. In dry fermentation, the coffee cherry fruits are dried and ditions and fermentation by Saccharomyces cerevisiae.
they remain intact. While, in Washed fermentation, coffee cherries are
hulled before drying and the mucilage is degraded by a mechanical or 2. Material and methods
microbial process (Oliveira, Biaggioni, Borém, Isquierdo, & Damasceno,
2018). In addition, the coffee industry and rural producers use differ- This study was carried out at the Laboratory of Analysis and Research
ent fermentation methodologies to alter the final quality of the coffee in Coffees (LAPC), of the Federal Institute of Espírito Santo (IFES) and
beverage. the Laboratory of Mycorrhiza Associations (LAMIC) of the Universidade
Fermentation has the potential to increase the scores for sensory at- Federal de Viçosa (Brazil).
tributes, such as flavor, texture, body, and acidity of the coffee bev-
erage when compared to beverages produced by unfermented coffee 2.1. Raw materials
(Elhalis, Cox, Frank, & Zhao, 2020, De Bruyn et al., 2017). These sensory
notes are mainly due to the influence of microorganisms on the coffee Arabica coffee fruits (Coffea arabica, Catuaí Vermelho variety) from
fruit (da Silva et al., 2022, Pereira et al., 2020). the 2019 harvest were used in this study. The harvest was performed
Spontaneous fermentation and/or induced fermentation may be used manually, with the cherry and greenish maturation degree fruits being
in coffee fermentation (Brioschi Junior et al., 2021, De Bruyn et al., selected. The washing and decanting of the fruits were carried out in
2017). Spontaneous fermentation is carried out with indigenous mi- a metallic container (capacity 200 L) with drinking water. The floating
croorganisms from the coffee fruit. This fermentation in coffee is a pro- fruits were discarded and the fruits with the highest densities (cher-
cess that occurs naturally in coffee fruits due to the number of free sug- ries and unripe fruits) were used in the Dry fermentation or Washed
ars and the richness of the microbiota (Pereira et al., 2020, Somporn, fermentation process. These processes were performed under anaerobic
Kamtuo, Theerakulpisut, & Siriamornpun, 2012). The microorganisms conditions using Saccharomyces cerevisiae (a commercial strain for bread
used in coffee fermentation can be present in the fruit itself (indigenous production).
microbiota) or in starter cultures (De Bruyn et al., 2017, De Melo, Soc- In dry fermentation, intact fruits (cherries and unripe fruits) were
col, Brar, Neto, & Soccol, 2017, Silva, Vilela, Cordeiro, Duarte, Dias, & used, while in Washed fermentation, the peel was removed, keeping
Schwan, 2013). the mucilage adhered to the surface of the parchment (da Silva Oliveira
The use of starter cultures is a method capable of obtaining better et al., 2022; Pereira et al., 2020). In both fermentations, the coffees fruits
sensory results during coffee fermentation, allowing for a more distinct were packed in polypropylene containers with a capacity of 20 L. One
flavor in the beverage than spontaneous fermentation (Ribeiro et al., percent (w/v) of S. cerevisiae with 107 Colony Forming Units per mL
2017). The isolation of indigenous microorganisms from the coffee fruit was inoculated into these containers with 20 kg of fruits or coffee beans
during post-harvest processing may be a viable alternative for obtain- and 10 L of water. After inoculation, the containers were hermetically
ing starter cultures. In isolated cultures, bacteria represent the major- closed with a plastic lid and incubated at 28°C.
ity of microorganisms, with the genera Arthrobacter, Klebsiella, Lacto- Fermentation times were 36, 72, 108, 144, 180, 216, 252, and 288
bacillus, Bacillus, Acinetobacter, and Weissella being commonly found; hours. Fruit sampling was performed at each fermentation time.
followed by yeasts, such as Saccharomyces and Rhodotorula and fila- The samples (0.5 kg) for microbial analyses were placed in transpar-
mentous fungi (Brioschi Junior et al., 2020; Ruta & Farcasanu, 2021; ent polyethylene bags (1 kg capacity) and stored at -20°C. For chemical
Vilela, Pereira, Silva, Batista, & Schwan, 2010). In a study carried out and sensorial analyses, there was no storage of the samples before the
with starter cultures of Saccharomyces cerevisiae, Candida parapsilosis, fruit drying process.
and Pichia guilliermondii it was possible to obtain sensory character-
istics of caramel, herbs, and fruits in the coffee beverage (Haile & 2.2. Drying the fruits after fermentation
Kang, 2019). Yeast inoculation can modify the sensory profile and in-
crease the coffee beverage score by up to 5 points (Mota, Batista, Rabelo, Coffee samples from different processing and fermentation times
Ribeiro, Borém, & Schwan, 2020). According to these authors, the in- were washed in potable water and dried until reaching a moisture con-
oculation of S. cerevisiae and Torulaspora delbrueckii are more suitable tent of 11% wet basis (Pereira et al., 2020). The coffees were sun-dried
for pulped natural coffee and natural coffee, respectively. Coffee beans for 15-18 days in a covered suspended system (plastic cover), with tem-
fermented for 24 hours with three different yeast strains showed an perature minimum (at night), average, and maximum of 7°C, 25°C, and
increase in antioxidant activity (Kwak, Jeong, & Kim, 2018). Further- 45°C, respectively. The Arduino Uno R3 system (Bluetooth Module Hc-
more, coffee beans fermented for 24 hours with three different yeast 06 Rs232- humidity and temperature sensor Dht22 Am2302 – SD card
strains showed an increase in antioxidant activity, polyphenol level, module) was used to monitor the temperature.
and flavonoid content when compared to unfermented coffee beans Before roasting, the samples were submitted to a parchment removal
(Kwak, et al., 2018). process to standardize the size of the beans using a 16-up sieve accord-
Microbial metabolism during fermentation causes changes in the ing to the recommendations of the sensory analysis protocol from the
chemical composition of coffee by the degradation of plant compounds Specialty Coffee Association (SCAA, 2013).
or by the synthesis of metabolites (Elhalis et al., 2020; De Bruyn
et al., 2017; Mota, Batista, Rabelo, Ribeiro, Borém, & Schwan, 2020; 2.3. Sample roasting and sensory analysis
Simmer et al., 2022). However, technical standardization of the pro-
cessing and fermentation steps is very important for food safety and to Coffee beans were roasted using a Probatino roaster and a set of
guarantee the quality of the beverage, since it prevents the growth of Agtron-SCAA disks. Roasting time varied between 8 and 12 minutes.
pathogenic and toxin-producing microorganisms. This standardization The charging temperature and final bean temperature were 170°C and
can also influence the choice of post-harvest processing and fermenta- 210°C, respectively. The roasting point (Medium roast) was determined
tion methodology best suited to the local physical-chemical and finan- by the color disks #65 and #55 that were used in the sensory analysis of
cial characteristics. Therefore, understanding the dynamics and com- specialty coffees. In this analysis, 5 cups per batch were prepared using
position of microorganisms during the fermentation of coffee fruits is of 8.25 grams of roasted and ground coffee (Debona, ten Caten, Guarçoni,
great importance to obtain a product with the desired chemical and sen- Moreira, Moreli, & Pereira, 2021; SCAA, 2013). The coffee samples were

2
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

ground using a Bunn G3 electric mill, with medium/coarse granulome-

try and 20 mesh patterns.
The panel of the sensory attributes (i.e. fragrance/aroma, flavor, af-
tertaste, sweetness, balance, acidity, body, uniformity, clean cup, and
final score) was evaluated by 9 Q-graders (SCAA, 2013). The total score
or final score on a scale of 0 to 100 percentage points corresponds to
the sum of the grades for each of these sensory attributes.

2.4. Mid-Infrared Analyses

The coffee samples were ground in a Bunn Coffee Mill (model G3A
HD), with granulometry between 70% and 75% of the particles passing
through a 0.074-mm mesh sieve.
An aliquot (10 g) of each sample (0.074 mm - mesh sieve, stan-
dard US Standards) was placed in the glass desiccator until it reached
room temperature (25 ± 2°C). For each sample, four pellets were
formed containing 3 g of ground coffee without the addition of KBr
Fig. 1. Final score of sensory attributes of coffee beverage fermented by Saccha-
(Tavares, Pereira, Nunes, Pinheiro, & Rodarte, 2012). romyces cerevisiae by dry fermentation and washed fermentation. The averages
The spectra were obtained in a model Cary 630 Fourier Transform In- followed by the same capital letter on the bars do not differ from each other by
frared (FTIR) spectrometer from the manufacturer Agilent Technologies, the Tukey test at 5% probability. Coffee classification according to Uganda Cof-
in an ATR diamond accessory with a reflection angle of 45°, 1 mm diam, fee Development Authority (2012).
200 𝜇m of the active area, and approximately 2 𝜇m of penetration depth
in the sample, using a Zinc selenide (ZnSe) reflectance detector. The
mid-infrared spectra of the samples were obtained as the average of 8
consecutive scans, with a resolution of 4 cm−1 in the wavenumber range adjustment of the sensory variables made by the function available in
from 4000 to 650 cm−1 . This region of the mid-infrared electromagnetic the vegan library 2.5.6. To measure the variation of microbial commu-
spectrum provides a large amount of information that can be used for nity between periods of fermentation, we calculated the distance be-
the functional characterization of organic compounds (Pavia, Lampman, tween the weighted vectors in the seven-dimensional space of the PcoA
& Kriz, 2014). (i.e. each dimension was multiplied by the proportional eigenvalue of
that dimension). High and low values mean, respectively, the high and
2.5. DNA extraction, PCR, and denaturing gradient gel electrophoresis low variance of the community.
The DDGE dendrograms were constructed using Dice’s similarity in-
Total DNA was extracted from 250 mg of fruits macerated in liq- dex and cluster analysis with Ward’s minimum variance method in the
uid nitrogen, followed by using a Nucleo Spin Soil Kit (MachereyNagel, Bionumeric software (version 6.0, Applied Maths, Inc., Austin, Texas,
GmbH & Co. KG, Germany), according to the manufacturer’s instruc- USA). The richness, evenness (Pielou), and diversity (Simpson) of mi-
tions. croorganisms were obtained in the vegan 2.5.6 package (Oksanen et al.,
The PCR and Denaturing Gradient Gel Electrophoresis (DGGE) were 2019) using the R software.
performed as described in Brioschi Junior et al. (2021) and Da Silva
et al. (2022) using the primers F27 and 1492R on 16S rDNA and ITS1 3. Results and discussion
and ITS4 on the ITS region of fungi.
3.1. Sensory quality
2.6. Experimental design and statistical analysis
The total score for fermented coffee (Fig. 1) was higher than for un-
The experiment was carried out in a completely randomized design fermented coffee from the same planting region (Debona et al., 2021).
with 2 types of post-harvest processing/fermentations, 8 fermentation An increase of up to 5 points in sensory scores of the coffee beverage
times, and 4 replications. Analysis of variance and Tukey’s test at 5% was observed after inoculation of Saccharomyces cerevisiae (CCMA 0543)
probability were performed for sensory data. and Torulaspora delbrueckii in Arabica coffee (Mota et al., 2020). How-
In the evaluation of similarities, tests were applied between the fer- ever, after 36 h of fermentation, there was no increase or decrease in
mentation times for the two processes, followed by the analysis under the overall beverage scores (Fig. 1). Da Silva et al. (2022) also did not
the perception of the Q-Graders about the fermentation methods. In ad- observe significant differences in the overall score of fermented coffee
dition, matrices with the averages of the variables and dendrograms with S. cerevisiae between 36 h and 72 h. Furthermore, S. cerevisiae may
with Euclidean distance to measure the distances between the fermen- act as a bioprotective agent in the fermentation process due to nutrient
tation times and the sensory score of the Q-Graders were performed. and oxygen consumption and alcohol synthesis that provides a selec-
The hierarchical complete linkage method was used for the observation tive fermentation environment for microbial growth (Ruggirello et al.,
of clusters. 2019). A reduction in the richness of the coffee microbial community
Multiplicative scatter correction (MSC) was applied to the spectral was observed in coffee fruits fermented with S. cerevisiae (Da Silva et al.,
data. The infrared spectra were organized in a matrix. In this matrix, 2022), which may be due to this selective environment. According to
each replica was considered a sample. The data were centered on the Elhalis et al. (2020), yeasts play a crucial role during the wet fermenta-
average and subjected to Principal Component Analysis (PCA) and Hi- tion of coffee beans for the production of highly sensory quality coffee.
erarchical Cluster Analysis (HCA) to detect similarities and differences According to the Uganda Coffee Development Authority classifica-
in chemical composition among the coffees analyzed. tion (UCDA, 2012), coffees fermented by dry and wet methods can be
All calculations (Euclidean distance, MSC, PCA, and HCA) were per- classified as fine coffee (Fig. 1). Thus, these fermentation processes el-
formed in Matlab software version R2013a and Microsoft Excel version evate the classification of Arabica coffees produced in the state of Es-
2016. pírito Santo (Brazil) from premium coffee to fine coffee. The final scores
The analysis of Principal Coordinates Analysis (PCoA) was per- of coffee beverage sensory attributes on dry and wet processing had a
formed based on the Bray-Curtis distance between the values and the positive linear effect with planting altitude. Furthermore, according to

3
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

Fig. 2. Dendrograms of the dry fermentation (A) and washed fermentation (B) of coffee at the fermentation times (36, 72, 108, 144,180, 216, 252, and 288 hours)
using Saccharomyces cerevisiae and the relative contribution (%) of sensory attributes for the construction of these dendrograms (C=dry fermentation and D=washed
fermentation).

Pereira et al. (2020), these positive effects of washed fermentation sup- significant differences in the total scores depending on the type and
port new sensory routes with increased coffee beverage flavor. There- time of fermentation, except for 108 and 216 h (Fig. 1). According
fore, coffee fermentation is important to increase the sensory quality of to Machado et al. (2021), the final score was not a good parame-
coffee. ter for sensory discrimination of the Coffea canephora genotypes due
In the fermentation times of 108 h and 216 h, the total score was to the disproportionality of its sensory scores when compared to the
higher in the washed fermentation than in the dry fermentation, which scores of the other sensory attributes of the beverage. Flavor, bitter-
also shows the influence of the type of processing after harvest and the ness/sweetness, aftertaste, and fragrance/aroma were the sensory at-
removal of the husk on the quality of fermented coffee by S. cerevisiae tributes with the greatest contribution to the overall note of the coffee
(Fig. 1). The linear correlation coefficient for the total score was lower beverage (Machado et al., 2021; Pereira et al., 2020). In addition, the
for dry fermentation than for washed method using coffees of differ- individual analysis of the sensory attributes (e.g. fragrance/aroma, acid-
ent plantation altitudes (Pereira et al., 2020), which shows the impacts ity, body, and balance) had different correlation coefficients with the
of the types of coffee-fruit processing on the quality of the beverage. composition of volatile compounds in Arabica coffee (Agnoletti et al.,
According to Mota et al. (2020), S. cerevisiae has a better metabolic per- 2022). Thus, to demonstrate the influence of processing type and fer-
formance in hulled coffee than in natural coffee. The fermentation time mentation time on coffee quality we made dendrograms using the Eu-
alters the profile of the coffee fungal and bacterial community, which clidean distance with the relative individual contributions of each sen-
may influence the synthesis of metabolites from microbial growths and sory attribute (Figs. 2-4, Supplementary material).
flavor-precursor compounds in fermented coffee (Brioschi Junior et al.,
2021). Furthermore, Da Silva et al. (2022) showed that fruit processing
3.2. Sensory grouping of coffees according to fermentation processes and
(in natura, pulp addition, pulp removal, and superficial disinfestation)
time
positively contributes to the diversity and richness of microorganisms
during fermentation and to the chemical and sensory quality of the cof-
Three homogeneous groups were observed in the dendrograms as a
fee beverage.
function of fermentation time and sensory attributes (Fig. 2A and B).
Long fermentation times (> 100 hours) can also cause a possible
These groups show the influence of time and type of fermentation on
loss of sensory quality due to catabolic processes. Acetic fermentation
the sensory attributes of the coffee beverage evaluated by the Q-Graders
occurs through the transformation of alcohol into acetic acid (Rainieri
(panelists).
& Zambonelli, 2009). This loss of quality could also be correlated to the
In dry fermentation, group A was formed with fermentation times
increase in ammonia concentration following fermentation since this
of 108h and 216h (Fig. 2A). Group B was formed by the fermentation
compound is liberated from the oxidative deamination of amino acids
time of 288h and group C by the other times (Fig. 2A). While in washed
(Lee, Geng, Mun, Curran, Bin, & Liu, 2016).
fermentation, group A (108 h and 216 h), group B (72, 180, and 252 h),
The total score is the sum of the individual scores of each sen-
and group C (36, 144, and 288 h) had a more equitable distribution of
sory attribute (Pereira et al., 2019), which may explain the lack of
fermentation times. (Fig. 2B).

4
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

Fig. 3. Dendrograms between the nine Q-Graders and their respective interactions at the times of 36, 72, 108, 144,180, 216, 252, and 288 hours in dry fermentation
and washed fermentation.

5
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

Fig. 4. Dendrogram between fermentation times (A) and (B), loading graph (C and D) for the dry fermentation (A and C) and washed fermentation (C and D) of the
infrared spectroscopy data of the selected regions.

The groups show that sensory attributes of the coffee beverage observed that this sensory attribute was lower in natural coffee than in
were influenced by the type and fermentation time of the coffee fruits washed coffee.
(Fig. 2C and D). Balance and body were the sensory attributes that most
contributed to the formation of the three groups in dry fermentation
3.3. Relationship between Q-graders and fermentation methods
(Fig. 2C). Flavor, acidity, overall, and total score had no contribution to
the formation of groups in dry fermentation. Meanwhile, in washed fer-
In the dendrograms, it was observed that the Q-Graders did not re-
mentation, the fragrance had the highest percentage of contribution to
main in homogeneous groups during the sensory analysis of the coffee
the formation of the groups that other sensory attributes and the flavor
beverage obtained at different times of dry fermentation (Fig. 3, Sup-
and the total note had no percentage of contribution (Fig. 2D). The high
plementary material).
percentage contributions of fragrance in the dendrograms (Fig. 2C and
The dendrograms of the washed fermentation suggest similar results
D) are in agreement with the observations of Lee et al. (2016) who re-
to the dry fermentation, where the Q-Graders did not form homogeneous
ported an improvement in the fragrance quality of fermented coffee with
groups throughout the fermentation periods (Fig. 3). These changes in
the use of Saccharomyces spp. The production of volatile compounds
the formation of sensory groups are a reflection of the high sensitiv-
such as acetaldehyde, ethanol, ethyl acetate, and isoamyl acetate pro-
ity that these evaluators have or the great subjectivity of the applica-
vides desirable aromatic characteristics such as caramel and fruitiness,
tion of the sensory analysis technique using the SCA protocol (Di Don-
in addition to typical attributes of yeast fermentation, such as buttery
francesco, Gutierrez Guzman, & Chambers, 2014). In addition, results
notes (Ruta & Farcasanu, 2021). Thus, the results of long-duration anaer-
by Schmidt and Miglioranza (2010) show that sensory analysis is a sub-
obic yeast fermentation indicate that it is possible to develop different
jective test, as it deals with human behavior. However, the studies by
sensory profiles, so that producers can develop fermentation strategies
Pereira et al. (2017, 2019, 2020) confirm that Q-Graders have a strong
to offer the specialty coffee market with differentiated beverages. An-
level of accuracy in the sensory evaluation of coffees with the SCA pro-
other relevant factor observed in washed fermented coffee was acidity
tocol. Evidence of changes in sensory groups is a reflection of the high
(Fig. 2D). Hameed, Hussain, Ijaz, Ullah, Pasha, & Suleria (2018) also
sensitivity of sensory evaluators (Pereira et al. 2019).

6
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

Fig. 5. Dendrogram following the DICE WARD analysis obtained from the DGGE banding pattern of the bacteria (A) and fungi (B) and biological biodiversity indices
(C) of coffee samples at different periods of fermentation (36, 72, 108, 144, 180, 216, 252, and 288 h).

In both fermentations (dry and washed), the contributions of the sen- such similarities or differences in the data (Fig. 4). Deference in clusters
sory attribute scores to the formation of dendrograms varied over the of dry fermentation and washed fermentation was observed in these
fermentation times (Fig. 3, Supplementary material). These alterations HCA analyses (Fig. 4A and B).
suggest a transition of the sensorial profile during the fermentation time For the dry fermentation, the PCA/HCA analyses suggest the forma-
by the perception of the Q-Graders used in this study. Thus, the discrim- tion of 3 groups of samples: group A formed by the fermentation times
ination of sensory profiles in fermentation times may be related to the of 72h, 144h, and 180h, group B by the fermentation time of 36h, and
diversity of microorganisms and the variation in the chemical composi- group C by the other times (Fig. 4 A). While in the washed fermentation,
tion of coffee fruits during fermentation. PCA/HCA shows the formation of two main groups: group A formed by
Time and temperature change the microbial profile during coffee fer- the fermentation times 144h, 180h, and 216h and group B by the other
mentation (Mota et al., 2020). According to Nakayama et al. (2020), the times (Fig. 4 B). The distribution of fermentation times is due to dif-
prevalence of a species and its microbial succession in the coffee bean ferences in the chemical composition of the samples (Fig. 4C and D).
depends on the physiochemical conditions of the environment and the The spectral regions with the greatest influence on this separation are
chemical composition of the fruit. This microbiological succession can 3600 to 3100 cm−1 , 3000 to 2850 cm−1 , and 1800 to 1750 cm−1 . In the
occur throughout cultivation in the field, the harvest, and the process- wavelength range of 3600 to 3100 cm−1 , vibrations of O-H bonds from
ing of the grains. Diversity of microorganisms such as Lactobacillus sp., axial deformation occur in hydrogen atoms bonded to nitrogen. The pri-
Hanseniaspora sp., Meyerozyma sp., and Debaryomyces sp. may be ob- mary amines, amides of amino acids and proteins, or O-H bonds from
served during coffee fermentation (Evangelista et al., 2015; Mota et al., in water are observed in this region. In the region from 3000 to 2850
2020). cm-1, elongation of the aliphatic C-H bond of caffeine can occur and the
The carbohydrates (pectin, cellulose, hexoses, and pentoses) of the region around 1750 cm−1 is related to the absorption of esters present
coffee fruit favor colonization by filamentous fungi, yeasts, and bacteria in lipids (Collazos-Escobar, Gutiérrez-Guzmán, Váquiro-Herrera, Bom,
that alter the composition of the coffee fruit by the synthesis of biomass Garcia-Perez, 2022; Pavia et al., 2014).
and metabolites and by enzymatic degradation (pectin-lyase, polygalac- The spectral bands of dry fermentation and washed fermentation
turonase, cellulose, and pectin-methyl-esterase) from plant compounds show a difference only in the region of 1100-1050 cm−1 , which may
(Haile & Kang, 2019; Pereira et al., 2020). Reducing sugars are impor- be associated with bands of secondary alcohols (Fig. 2CD). These com-
tant flavor precursor in coffee (Somporn et al., 2012). Furthermore, pounds in washed fermentation (Fig. 2D) may be due to the fermentation
in fermentation, microorganisms also release a variety of compounds, of sugars after husk removal, as also observed in other studies of bev-
such as organic acids, esters, and alcohols that directly affect the sen- erages (da Silva Oliveira et al., 2022; Elhalis et al., 2020; Pereira et al.,
sory characteristics of coffee (Martins, Batista, Miguel, Simão, Soares, & 2020).
Schwan, 2020).
3.5. Microbial community profile
3.4. Grouping of coffees according to fermentations (dry and washed) and
fermentation time based on the chemical composition For the bacterial community, two main groups with 20% similar-
ity were observed in the DGGE profiles (Fig. 5 A). The first group with
The PCA and HCA were applied to all spectra (4000-650 cm−1 ) and 36, 72, 108, and 288 h of fermentation, that is, the shortest times com-
the regions without peaks were removed, to better understand the data bined with the longest fermentation time with 40% similarity. Also, it
patterns and to find clusters of samples and spectral peaks by directing is noteworthy that at 108 h of fermentation, there was no difference re-

7
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

Fig. 6. (A) - Principal Coordinates analysis (PCoA) of bacterial and fungal communities based on Bray-Curtis dissimilarity, fermented coffee by dry and washed
methods at the different fermentation times (36, 72, 108, 144, 180, 216, 252 and 288 h). The sensorial variables were fitted to the ordination using the envfit function
of the vegan package in R. (B) - Variation of communities along the fermentation process.

garding the type of fermented coffee, showing 100% similarity. At 36 and pulped coffee and the 144-h sample of natural coffee (Fig. 5 B). The
and 72 h the community present in the pulped or natural coffee showed third group with almost 100% similarity, grouped mainly samples of
about 90% similarity (Fig. 5 A). The results show that the coffee peeling pulped coffee from intermediate fermentation times (108, 144,180, and
does not cause a major change in the community of bacteria. The sec- 216 h). This shows that peeling the fruit affects the community, differ-
ond group was formed by the fermentation times of 144, 180, 216, and entiating it from the natural coffee community. The formation of groups
252 h and showed 50% similarity (Fig. 5 A). At the 252-h fermentation in the PCR-DGGE with different percentages of similarity was observed
time, the type of coffee bean used (pulped or natural coffee) showed in coffees fermented by washed and dry methods using 28S rDNA for
90% similarity. In group 2, natural coffee fermented for 144 hours was fungi and 26S rDNA for yeasts (Durand et al., 2013). According to these
the most distinct, with approximately 45% similarity. The PCR-DGGE authors, there was a difference in PCR-DGGE profiles between fermen-
profile, CFU/g, and diversity of bacteria and yeast were different every tation methods and PCR-DGGE is efficient to evaluate the dynamics of
20 hours of fermentation of coffee fruits using wet and dry methods for the fungal community before and during the fermentation of the coffee
360 h (Evangelista et al., 2015). Thus, there is a change in the coffee fruits. Furthermore, microbial ecology structures depend mainly on the
fruit bacterial community with fermentation time, which may be influ- fermentation method, the coffee species, and the geographical origin of
enced by the chemical and sensory profiles of coffee beverages (Figs. 5 the coffee (Hamdouche et al., 2016).
and 6). The communities of bacteria in both treatments, dry and washed
The profile of the fungal community showed three groups with 40% fermentation, showed a decrease in the richness along the 288 h of fer-
similarity (Fig. 5 B). The first was formed by samples with longer fer- mentation (Fig. 5 C). The washed treatment, for instance, decreased the
mentation times (216, 252, and 288 h) and the naturally fermented cof- richness by half (16 to 8). On the other hand, the fungal community
fee (Fig. 5 B). The second group with 80% similarity was composed of showed a slight variation along the 288 h of fermentation. The decrease
samples with short fermentation times (36 and 72 h) of both natural in bacterial richness, however, was followed by an increase in the even-

8
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

ness of the bacterial community. This shows that the bacterial commu- CRediT authorship contribution statement
nity tends to be more homogeneous after a long time of fermentation
(288 h), with less dominance of some species over others. According to Luiz Henrique Bozzi Pimenta de Sousa: Conceptualization,
Evangelista et al. (2015), bacterial growth is favored in wet processing Methodology. José Maria Rodrigues da Luz: Conceptualization,
compared to dry fermentation. However, the impact of adding water to Formal analysis, Methodology, Writing – original draft, Writing –
coffee fermentation was dependent on time with alternating dominance review & editing. Marliane de Cássia Soares da Silva: Conceptual-
between the growth of bacteria and fungi throughout the fermentation ization, Formal analysis, Methodology, Supervision, Writing – original
times (Da Silva et al., 2022). Thus, the method and time of fermenta- draft, Writing – review & editing. Aldemar Polonini Moreli: Con-
tion influence the microbial communities (Figs. 5 and 6). According to ceptualization, Formal analysis, Methodology, Supervision, Writing
Da Silva et al. (2022), the microbial profile and sensory score of coffee – review & editing. Tomás Gomes Reis Veloso: Conceptualization,
beverages are dependent on the conditions of processing of coffee fruits Formal analysis, Methodology, Writing – original draft, Writing –
and fermentation. review & editing. Rogério Carvalho Guarçoni: Conceptualization,
The Principal Components Analysis (PCoA) showed that for both bac- Methodology. Taís Rizzo Moreira: Conceptualization, Formal analysis,
teria and fungi the final score, body, and acidity tend to be higher in Methodology. Marcos Vinícius Pereira Barros: Conceptualization,
samples fermented for a longer time (Fig. 6 A). In general, the fungal Formal analysis, Methodology. Maria Catarina Megumi Kasuya: Con-
community composition was more variable along the fermentation pro- ceptualization, Formal analysis, Supervision, Writing – original draft,
cess than the bacterial community (Fig. 6 B). The bacterial community Writing – review & editing. João Paulo Pereira Marcate: Conceptual-
showed a considerable change in the final stage of the fermentation pro- ization, Formal analysis, Methodology, Supervision, Writing – review
cess. The pulped coffee showed a high change between 252 and 288 & editing. Dério Brioschi Júnior: Conceptualization, Formal analysis,
h. This high change in the bacterial community composition coincided Writing – original draft, Writing – review & editing. Willian dos Santos
with the best final score of the pulped coffee. The dry fermentation, in Gomes: Conceptualization, Formal analysis, Writing – original draft,
turn, had the most pronounced changes from 216 to 288 h (Fig. 6 B). Writing – review & editing. Lucas Louzada Pereira: Conceptualization,
The fungal community showed more variability along the fermentation Formal analysis, Methodology, Supervision, Writing – original draft,
process than the bacterial community, with the changes in the dry fer- Writing – review & editing. Emanuele Catarina da Silva Oliveira:
mentation more pronounced (Fig. 6 B). In general, as observed in Fig. 5, Conceptualization, Formal analysis, Methodology, Supervision, Writing
the bacterial communities of pulped and natural coffee with the same – original draft, Writing – review & editing.
fermentation time tend to be more similar to each other than the fungal
community. According to Pereira et al. (2020), the interactions between Data availability
edaphoclimatic factors, forms of processing, drying, storage, and roast-
ing, directly affect the sensorial attributes of coffee beverages. Our stud- Data will be made available on request.
ies using samples of different planting regions have shown the positive
impacts of the microbiota of the fruits and soils of the coffee plant on Acknowledgments
beverage quality (Da Silva Oliveira et al., 2022; Simmer et al., 2022).
The production of primary and secondary metabolites and the degrada- The authors thank the Sul Serrana of Espírito Santo Free Admission
tion of mucilage are the main microbial activities observed during the Credit Cooperative – Sicoob (23186000886201801), CAPES (Coorde-
fermentation of the coffee fruit (De Bruyn et al., 2017). Furthermore, nação de Aperfeiçoamento de Pessoal de Nível Superior – Código de Fi-
crop management and growing environments may be more important nanciamento 001), CNPq (Conselho Nacional de Desenvolvimento Cien-
determinants of the quality of Coffea arabica bean than genetic factors tífico e Tecnologia – Código 304087/2020-3), EMBRAPA, INCAPER, and
(Bote & Vos, 2017). Instituto Federal do Espírito Santo, for supporting the research through
the PRPPG n°. 10/2019 – Productivity Researcher Program – PPP. Our
4. Conclusions thanks are also extended to the rural producer José Luiz Pimenta, for
donating the coffee to the study, and to the researchers of the Coffee
The induced fermentation of coffee by Saccharomyces cerevisiae in- Designer Group - dgp.cnpq.br/dgp/espelhogrupo/5617656954660266.
creases the sensory classification of the coffee beverage with positive ef-
fects on the chemical and sensory profile. These positive impacts depend Supplementary materials
on the type (dry or washed) and time of fermentation that contribute to
microbial growth in the fruit. Natural coffee (dry fermentation) had the Supplementary material associated with this article can be found, in
greatest sensory profile complexity in fermented coffee with yeast in an the online version, at doi:10.1016/j.focha.2023.100259.
anaerobic environment and the maintenance of quality extends accord-
ing to the fermentation time. Furthermore, our results are important for References
understanding the fermentation processes of coffee beans and choosing
the most appropriate type and fermentation time for the development Agnoletti, B. Z., Folli, G. S., Pereira, L. L., Pinheiro, P. F., Guarçoni, R. C.,
of specialty coffees. Oliveira, E. C. S., & Filgueiras, P. R. (2022). Multivariate calibration applied to study
of volatile predictors of arabica coffee quality. Food Chemistry, 367, Article 130679.
10.1016/j.foodchem.2021.130679.
Bertrand, B., Boulanger, R., Dussert, S., Ribeyre, F., Berthiot, L., Descroix, F., &
Supplementary material
Joët, T. (2012). Climatic factors directly impact the volatile organic compound finger-
print in green arabica coffee bean as well as coffee beverage quality. Food Chemistry,
Percentage of contribution of sensory attributes in the construction 135(4), 2575–2583. 10.1016/j.foodchem.2012.06.060.
of dendrograms for washed fermentation (fig. 3) and dry fermentation Bhumiratana, N., Koushik, A., & Chambers, E. (2011). Evolution of sensory aroma at-
tributes from coffee beans to brewed coffee. LWT - Food Science and Technology,
(fig. 4) as a function of fermentation time. 44(10), 2185–2192. 10.1016/j.lwt.2011.07.001.
Bote, A. D., & Vos, J. (2017). Tree management and environmental conditions affect coffee
(Coffea arabica L.) bean quality. NJAS - Wageningen Journal of Life Sciences, 83, 39–46.
Declaration of Competing Interest 10.1016/j.njas.2017.09.002.
Brioschi Junior, D., Guarçoni, R. C., Da Silva, M. C. S., Veloso, T. G. R., Kasuya, M. C. M.,
The authors declare that they have no known competing financial da Silva Oliveira, E. C., da Luz, J. M. R., Moreira, T. R., Debona, D. G., &
Pereira, LL. (2021). Microbial fermentation affects sensorial, chemical, and micro-
interests or personal relationships that could have appeared to influence bial profile of coffee under carbonic maceration. Food Chemistry, 342, Article 128296.
the work reported in this paper. 10.1016/j.foodchem.2020.128296.

9
L.H.B.P. de Sousa, J.M.R. da Luz, M.d.C.S. da Silva et al. Food Chemistry Advances 2 (2023) 100259

Castellanos-Onorio, O., Gonzalez-Rios, O., Guyot, B., Tachon, A., Fontana, J., Guiraud, P., Mota, M. C. B., Batista, N. N., Rabelo, M. H. S., Ribeiro, D. E., Borém, F. M., &
Schorr-Galindo, S., Durand, N., & Suárez-Quiroz, M. (2011). Effect of Two Different Schwan, R. F. (2020). Influence of fermentation conditions on the sensorial qual-
Roasting Techniques on the Ochratoxin A (OTA) Reduction in Coffee Beans (Coffea ity of coffee inoculated with yeast. Food Research International, 136, Article 109482.
Arabica).”. Food Control, 22, 1184–1188. 10.1016/j.foodcont.2011.01.014. 10.1016/j.foodres.2020.109482.
Collazos-Escobar, G. A., Gutiérrez-Guzmán, N., Váquiro-Herrera, H. A., Bom, J., & Garcia- Nakayama, C. C., Teixeira, A. A., Teixeira, R. R., Reis, M., Monteiro, A., Bueno, J., &
Perez, J. V. (2022). Thermodynamic analysis and modeling of water vapor adsorption Taniwaki, M. H. (2020). Microorganism succession at different coffee drying stages
isotherms of roasted specialty coffee (Coffee arabica L. cv. Colombia). LWT - Food and its influence on the beverage. Brazilian Journal of Development, 6(1), 2402–2418.
Science and Technology, 160, Article 113335. 10.1016/j.lwt.2022.113335. 10.34117/bjdv6n1-177.
da Silva Oliveira, E. C., da Luz, J. M. R., de Castro, M. G., Filgueiras, P. R., Guarçoni, R. C., Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., Mcglinn, D.,
Castro, E. V. R., Silva, M. C. S., & Pereira, L. L. (2022). Chemical and sensory discrim- Minchin, P. R., O’hara, R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H.,
ination of coffee: Impacts of the planting altitude and fermentation. European Food Szoecs, E. H., & Wagner, H. (2019). Vegan: Community ecology package. R package
Research and Technology, 248, 659–669. 10.1007/s00217-021-03912-w. version 2.5-6.
da Silva, M. C. S., da Luz, J. M. R., Veloso, T. G. R., Gomes, W. S., Oliveira, E. C. S., Oliveira, P. D., Biaggioni, M. A. M., Borém, F. M., Isquierdo, E. P., & Damas-
Anastácio, L. M., Cunha Neto, A., Moreli, A. P., Guarçoni, R. C., Kasuya, M. C. M., ceno, M. O. V. (2018). Quality of natural and pulped coffee as a function of tem-
& Pereira, L. L. (2022). Processing techniques and microbial fermentation on micro- perature changes during mechanical drying. Coffee Science, 13(4), 415–425.
bial profile and chemical and sensory quality of the coffee beverage. European Food Pavia, D. L., Lampman, G. M., & Kriz, G. S. (2014). Introduction to spectroscopy (p. 784).
Research and Technology, 248, 1499–1512. 10.1007/s00217-022-03980-6. Brazilian, Brazil: Cengage Learning.
De Bruyn, F., Zhang, S. J., Pothakos, V., Torres, J., Lambot, C., Moroni, A. V., & De Pereira, L. L., Guarçoni, R. C., Pinheiro, P. F., Osório, V. M., Pinheiro, C. A., Moreira, T. R.,
Vuyst, L. (2017). Exploring the impacts of postharvest processing on the microbiota & ten Caten, C. S. (2020). New propositions about coffee wet processing: Chemical
and metabolite profiles during green coffee bean production. Applied and Environmen- and sensory perspectives. Food Chemistry, 310. 10.1016/j.foodchem.2019.125943.
tal Microbiology Journal, 83(1), e02398 16. 10.1128/AEM.02398-16. Pereira, L. L., Cardoso, W. S., Guarçoni, R. C., Fonseca, A. F. A., Moreira, T. R.,
De Melo, G. V. P., Soccol, V. T., Brar, S. K., Neto, E., & Soccol, C. R. (2017). Microbial & ten Caten, C. S. (2017). The consistency in the sensory analysis of cof-
ecology and starter culture technology in coffee processing. Critical Reviews in Food fees using Q-Graders. European Food Research and Technology, 243, 1545–1554.
Science and Nutrition, 57(13), 2775–2788. 10.1080/10408398.2015.1067759. 10.1007/s00217-017-2863-9.
Debona, D. G., ten Caten, C. S., Guarçoni, R. C., Moreira, T. R., Moreli, A. P., Pereira, L. L., Guarçoni, R. C., Moreira, T. R., Pimenta, L. H. B., & De Sousa Car-
& Pereira, L. L. (2021). Sensory analysis and mid-infrared spectroscopy for doso, W. S. (2019). Very beyond subjectivity : The limit of accuracy of Q-Graders.
discriminating roasted specialty coffees. Coffee Science, 16, Article e161878. Journal of Texture Studies, 50(2), 172–184. 10.1111/jtxs.12390.
10.25186/.v16i.1878. Rainieri, S., & Zambonelli, C. (2009). Organisms associated with acetic acid bacteria in
Di Donfrancesco, B., Gutierrez Guzman, N., & Chambers, E. (2014). Comparison of results vinegar production. In L. Solieri, & P. Giudici (Eds.), Vinegars of the world. Milano:
from cupping and descriptive sensory analysis of Colombian brewed coffee. Journal Springer. 10.1007/978-88-470-0866-3_5.
of Sensory Studies, 29, 301–311. 10.1111/joss.12104. Ribeiro, L. S., Ribeiro, D. E., Evangelista, S. R., Miguel, M. G. C. P., Pinheiro, A. C. M.,
Duguma, T. F., & Meer, P. V. (2018). Value chain analysis of the pre and Borém, F. M., & Schwan, R. F. (2017). Controlled fermentation of semi-dry coffee
post-harvest factors deteriorating the quality of coffee in the chole district, (Coffea arabica) using starter cultures: A sensory perspective. LWT - Food Science and
oromia region, Ethiopia. Food Science and Quality Management, 71, 18–23. Technology, 82, 32–38. 10.1016/j.lwt.2017.04.008.
https://iiste.org/Journals/index.php/FSQM/article/view/40629. Ruggirello, M., Nucera, D., Cannoni, M., Peraino, A., Rosso, F., Fontana, M., Co-
Durand, N., El Sheikha, A. F., Suarez-Quiros, M. L., Oscar, G. R., Nganou, N. D., Fontana- colin, L., & Dolci, P. (2019). Antifungal activity of yeasts and lactic acid bacteria
Tachon, A., & Montet, D. (2013). Application of PCR-DGGE to the study of dynamics isolated from cocoa bean fermentations. Food Research International, 115, 519–525.
and biodiversity of yeasts and potentially OTA producing fungi during coffee process- 10.1016/j.foodres.2018.10.002.
ing. Food Control, 34(2), 466–471. 10.1016/j.foodcont.2013.05.017. Ruta, L. L., & Farcasanu, I. C. (2021). Coffee and yeasts: From flavor to biotechnology.
Elhalis, H., Cox, J., Frank, D., & Zhao, J. (2020). The crucial role of yeasts in the wet Fermentation, 7(9). 10.3390/fermentation7010009.
fermentation of coffee beans and quality. International Journal of Food Microbiology, Specialty Coffee Association of America (pp. 1–10). (2013).
333. 10.1016/j.ijfoodmicro.2020.108796. Schmidt, C. A. P., & Miglioranza, E. (2010). The sensory analysis and the coffee: An revi-
Evangelista, S., Ferreira, M. P., Marques, C. A., & Freitas, R. (2015). Microbiological diver- sion (in Portuguese). Revista Eletrônica Científica Inovação e Tecnologia, 1, 2175. 1846
sity associated with the spontaneous wet method of coffee fermentation. International https://periodicos.utfpr.edu.br/recit/article/view/4096 .
Journal of Food Microbiology, 210, 102–112. Silva, C. F., Vilela, D. M., de Souza Cordeiro, C., Duarte, W. F., Dias, D. R., &
Haile, M., & Kang, W. H. (2019). The role of microbes in coffee fermenta- Schwan, R. F. (2013). Evaluation of a potential starter culture for enhance quality
tion and their impact on coffee quality. Journal of Food Quality, 7, 1–6. of coffee fermentation. World Journal of Microbiology and Biotechnology, 29(2), 235–
https://www.hindawi.com/journals/jfq/2019/4836709/. 247. 10.1007/s11274-012-1175-2.
Hamdouche, Y., Meile, J. C., Nganou, D. N., Durand, N., Teyssier, C., & Montet, D. (2016). Silva, S., Queiroz, D. M., Ferreira, W. P. M., Corrêa, P. C., & Rufino, J. L. S. (2016).
Discrimination of post-harvest coffee processing methods by microbial ecology anal- Mapping the potential beverage quality of coffee produced in the Zona Da Mata, Mi-
yses. Food Control, 65, 112–120. 10.1016/j.foodcont.2016.01.022. nas Gerais, Brazil. Journal of the Science of Food and Agriculture, 96(9), 3098–3108.
Hameed, A., Hussain, S. A., Ijaz, M. U., Ullah, S., Pasha, I., & Suleria, H. A. R. (2018). Farm 10.1002/jsfa.7485.
to consumer: Factors affecting the organoleptic characteristics of coffee. II: Posthar- Simmer, M. M. B., Soares da Silva, M., Pereira, L. L., Kasuya, M. C. M., Luz, J. M. R., & Silva
vest processing factors. Comprehensive Reviews in Food Science and Food Safety, 17, Oliveira, E. C. (2022). Edaphoclimatic conditions and the soil and fruit microbiota
1184–1237. 10.1111/1541-4337.12365. influence on the chemical and sensory quality of the coffee beverage. European Food
Kwak, H. S., Jeong, Y., & Kim, M. (2018). Effect of yeast fermentation of green coffee Research and Technology. 10.1007/s00217-022-04102-y.
beans on antioxidant activity and consumer acceptability. Journal of Food Quality, Somporn, C., Kamtuo, A., Theerakulpisut, P., & Siriamornpun, S. (2012). Effect of shad-
Article 5967130. 10.1155/2018/5967130. ing on yield, sugar content, phenolic acids and antioxidant property of coffee beans
Lee, L. W., Geng, Y. T., Mun, W. C., Curran, P., Bin, Y., & Liu, S. Q. (2016). Modulation of (Coffea Arabica L. cv. Catimor) harvested from north-eastern Thailand. Journal of the
coffee aroma via the fermentation of green coffee beans with Rhizopus oligosporus: I. Science of Food and Agriculture, 92, 1956–1963. 10.1002/jsfa.5568.
Green Coffee. Food Chemistry, 211, 916–924. 10.1016/j.foodchem.2016.05.076. Tavares, K. M., Pereira, R. G. F. A., Nunes, C. A., Pinheiro, A. C. M., & Ro-
Machado, J. L., Tomaz, M. A., da Luz, J. M. R., Osório, V. M., Costa, A. V., Colodetti, T. V., darte, M. P. (2012). Mid-infrared spectroscopy and sensory analysis applied to de-
Debona, D. G., & Pereira, L. L. (2021). Evaluation of genetic divergence of coffee tection of adulteration in roasted coffee by addition of coffee husks. Química Nova,
genotypes using the volatile compounds and sensory attributes profile. Journal of 35(6), 1164–1168. 10.1590/S0100-40422012000600018.
Food Science, 87(1), 383–395. 10.1111/1750-3841.15986. UCDA. (2012). Uganda coffee development authority Annual Report
Martins, P. M. M., Batista, N. N., Miguel, M. G. CP., Simão, J. B. P., Soares, J. R., & https://ugandacoffee.go.ug/resource-center/publications .
Schwan, R. F (2020). Coffee growing altitude influences the microbiota, chemical Vilela, D. M., Pereira, G. V., Silva, C. F., Batista, L. R., & Schwan, R. F. (2010). Molec-
compounds and the quality of fermented coffees. Food Research International, 129, ular ecology and polyphasic characterization of the microbiota associated with
Article 108872. semi-dry processed coffee (Coffea arabica L.). Food Microbiology, 27(8), 1128–1135.
Mota, M. C. B., Batista, N. N., Rabelo, M. H. S., Ribeiro, D. E., Borém, F. M., & 10.1016/j.fm.2010.07.024.
Schwan, R. F. (2020). Influence of fermentation conditions on the sensorial qual-
ity of coffee inoculated with yeast. Food Research International, 136, Article 109482.
10.1016/j.foodres.2020.109482.

10