Gonzalez Orozcoetal2023

Research Article
Received: 22 March 2023 Revised: 25 September 2023 Accepted article published: 27 September 2023 Published online in Wiley Online Library:
(wileyonlinelibrary.com) DOI 10.1002/jsfa.13016
Integrating new variables into a framework to

support cacao denomination of origin: a case
study in Southwest Colombia
Carlos E González-Orozco,a* Mario Porcel,a,b Roxana Yockteng,c,d
Alejandro Caro-Quintero,c,e Caren Rodriguez-Medina,f,g
Margareth Santander,c,g Martha Zuluaga,c Mauricio Soto,c
Jader Rodriguez Cortina,c Fabrice Eric Vaillantf
and Sebastian Escobar Parraf,g*
Abstract
BACKGROUND: Cocoa quality plays a pivotal role in establishing denominations of origin, with genotypes, geography, climate
and soil conditions being key variables. However, these factors have not been comprehensively explored in defining cacao
denominations of origin. The present study addresses this gap by laying the foundation for cacao denomination of origin, focus-
ing on the Buenaventura region on Colombia's Pacific coast. Our goal is to provide a holistic understanding of the elements under-
pinning cacao denomination of origin, emphasizing Buenaventura's unique cocoa quality and geographical significance.
RESULTS: Through the Buenaventura case, we propose a robust framework applicable to other cacao-producing regions, elevat-
ing the recognition and value of cacao denomination of origin. Our framework encompasses geography, agronomy, genetics,
microbial diversity, pests and diseases and cocoa quality. In a pioneering move, we propose a cacao denomination of origin in
Colombia, specifically examining Bajo Calima, Sabaletas and Cisneros within Buenaventura region. Buenaventura stands out
for its cocoa quality, characterized by fruity flavors attributed to the rich biodiversity of the lowland rainforest.
CONCLUSION: Our analysis indicates specific geographical indicators for each of the study zones, with Buenaventura identified
as a region with natural characteristics to produce fine flavour cocoa products. Each zone exhibited a high differentiation and
diversity of cacao cultivars. Buenaventura has the potential to be designated as a future denomination of origin for cacao from
the Pacific region of Colombia, characterized by its unique fruity-aroma chocolates. Our framework is adaptable to other cacao-
producing regions, facilitating the establishment of denominations of origin within the cocoa industry and agriculture.
© 2023 The Authors. Journal of The Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of
Chemical Industry.
Supporting information may be found in the online version of this article.
Keywords: Buenaventura; chocolate; cocoa flavour-aroma; rainforest
* Correspondence to: C E González-Orozco, Corporación Colombiana de Investigación Agropecuaria – Agrosavia. Centro de Investigación La Libertad- Km 14 vía
Villavicencio-Puerto López, Meta, Colombia. E-mail: [email protected]; or S Escobar, E-mail: [email protected]
a Corporación Colombiana de Investigación Agropecuaria – Agrosavia, Centro de Investigación La Libertad, km 14 via Puerto Lopez, VILLAVICENCIO, Meta, Colombia
b Instituto de Investigación y Formación Agraria, Pesquera, Alimentaria y de la Producción Ecológica (IFAPA), Málaga, Spain
c Corporación Colombiana de Investigación Agropecuaria – Agrosavia, Centro de Investigación Tibaitatá, vía a Mosquera, Bogotá, Colombia
d Departamento de Biología, Facultad de Ciencias, Universidad Nacional de Colombia sede Bogotá, Ciudad Universitaria, Bogotá, Colombia
e Corporación Colombiana de Investigación Agropecuaria – Agrosavia, Centro de Investigación La Selva, via Rionegro - Las Palmas, Sector Llano Grande, Rionegro,
Colombia
f Muséum National d'Histoire Naturelle, UMR-CNRS 7205, Paris, France
g Corporación Colombiana de Investigación Agropecuaria – Agrosavia, Centro de Investigación Palmira, Valle del Cauca, Colombia
1
© 2023 The Authors. Journal of The Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and
distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
10970010, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jsfa.13016 by Cochrane Colombia, Wiley Online Library on [19/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.soci.org CE González-Orozco et al.
INTRODUCTION of Cacao DOs in the tropical region of South America, a lack of sci-
Geographical indications (GI) and denomination of origin (DO) are entific rigor in the processes involved in supporting DO assign-
widely used concepts to determine the geographical and commer- ments is a common characteristic of all the mentioned cacao
cial status of food production.1 According to the World Intellectual cases. The documents supporting the request for the DO license
Property Organization,2 a GI is a sign used on goods with a specific often consist of four sections: (1) a legal framework; (2) a descrip-
geographical origin and possessing qualities or a reputation tion of the geographic region; (3) quality characteristics of the
because of that origin. DO, on the other hand, refers to food prod- product and its marketplace; and (4) communities and cacao asso-
ucts specific to a particular region or town, conveying a special ciativity or trading networks. Adopting a more evidence-based
quality or characteristic of the designated area (European Union and focused approach would greatly benefit intellectual property
Commission). Although the guidelines for GI and DO are globally and legal processes in the domain of food product qualifications.
accepted. standardized scientific frameworks to support their appli- Colombia lags behind in the realm of cacao DOs. However, it
cation to emerging products in developing economies are lacking. holds the potential to become a future stronghold for cacao in
Consequently, evidence-based frameworks are needed to provide the region as a result of its unique geographical conditions and
rigor in defining or understanding GI and DO. The present study diverse environments, encompassing 54 cacao regions in
presents a new approach to define and support the establishment Colombia.11 As mentioned before,12 achieving a comprehensive
of cacao GI and DO in an understudied tropical region. We propose system to improve GI and DOs legislation in Colombia requires
a science-based framework to support GI and DO establishment, significant inter-institutional coordination between the govern-
using the cacao-producing Pacific region of Buenaventura in ment, communities and private industry. Indeed, over the last
Colombia as a case study. Here, we will refer to the cacao tree as decade, the CCI has only granted 19.13 There are some examples
‘cacao’ and derived products made from cocoa beans as ‘cocoa’. of DOs in the south American coffee industry including
GI and DO are effective tools for differentiating foods and agro- Colombia and Honduras.14,15 However, the lack of scientific
nomical products.3 Global guidelines on GI regulations have been advances in cacao DOs and GIs in Colombia is counterproductive,
developed by FAO to strengthen capacities at various levels of the considering the country's abundant agrobiodiversity, and its sta-
food industry. Even within Europe, each country develops its own tus as a centre of cacao origin, accompanied by cultural richness.
regulations. In 2017, France published a booklet with user guide- Together, these factors present numerous opportunities in the
lines for GI and DO,4 providing clarity on rules for licensing DOs. In food industry. Despite these natural advantages, Colombia has
Latin America, the commission of the Andean community created yet to issue any cacao DOs or GIs. As mentioned in Section 3.3
a policy document on industrial property called Decision of the CCI technical manual, one of the weaknesses is limited sci-
486, which is related to patents of genetic resources. However, it entific research on the topic, compounded by a lack of qualified
does not offer a detailed guideline on DO requirements specifi- professionals in the field. However, the situation is currently evolv-
cally; instead, it states that registration is a matter for each coun- ing in Colombia. Recent cacao studies provide new evidence to
try. Consequently, regional legislations are unclear, complicating advance the creation of GIs for cacao in the country, such as pro-
the application process for GIs or DOs. This lack of clarity leads posing a cacao regionalization.11 This regionalization is the first of
to misinformation and a poor understanding of the commercial its kind and serves as the scientific foundation for defining the
value of GIs and DOs as tools. Colombia, for example, is no excep- separation of geographical cacao regions, which is useful because
tion. The Colombian Commerce and Industry institution (CCI), the it goes beyond political boundaries. It represents a novel
government entity responsible for issuing DOs, provides just four approach because the territory's geography drives the proposed
bullet points outlining the requirements for applying for a certifi- classification the river systems at different levels (i.e. catchment,
cate of denomination of origin. This list is clearly insufficient for sub-catchment), with microclimate based on the soil temperature
the extensive task at hand in the cocoa industry. and macroclimate data being used to define the cacao regions.
Native tropical cacao trees and their cultivated cousins are signif- This stands in contrast to existing cacao crop classifications, which
icant sources of cocoa aromas.5 However, they are not utilized as fre- rely on agronomical aptitude and departmental classification.16
quently as expected for proposing DOs and GIs. Several examples of The knowledge of cacao bean flavor quality serves as an exam-
cacao DOs or GIs exist in South America, such as Cacao Chuau in ple to illustrate the differentiation of cacao origins.17 Similar to
Venezuela and Ecuador, Colombia also produces fine-flavored
Venezuela,6 Cacao Arriba in Ecuador,7,8 Cacao Grijalva in Mexico9
cocoa. Certain Colombian regions have gained worldwide recog-
and Cacao Tomé-açu in Brazil.10 The Cacao Arriba region in
nition for their high-quality cocoa, solely based on organoleptic
Ecuador shares similar latitudes and climates with the Buenaventura
characteristics. However, there is a lack of scientific data support-
region. In Ecuador, the provinces of Los Rios and Guayas have a long
ing the holistic evaluation of Colombian cacao beans. Neverthe-
history as cacao-producing areas since the colonial period. In 2011,
less, recent studies have demonstrated the unique sensory
the Ecuadorian Institute of Intellectual Property granted a DO for qualities of Colombian cacao.18,19 Cocoa beans from the Arauca
Ecuador's fine aroma cacao, including the Cacao Arriba region on region in eastern Colombia have twice been awarded the title of
the Pacific Coast. However, the DO for Ecuador is applied under a best cacao in the world at the ‘International Cacao Awards’ held
broad scheme, and more specificity is needed in delimitating DO during the ‘Salon du Chocolat’ in Paris. This recognition was also
regions. For example, it would be an improvement if Ecuador could achieved by cacao beans from Tumaco region, situated in south-
derive from the National cacao a set of specific DO´s delimitated western Colombia geographically close to the Buenaventura
based on new surveys or methodologies. In the case of Venezuela, region. Additionally, other studies emphasized the quality poten-
the Cacao Chuau in Venezuela's Aragua state covers a specific geo- tial of the FEAR 5 cacao cultivar developed in the Arauca region,
graphic valley along a river that leads to the Caribbean Ocean. known by its fruity and nutty aromas.19 In parallel, more work
Validating DOs over time is a constant challenge for these has been carried out on developing a fermentation methodology
regions because Chuau's DO is still under scrutiny, and finalizing for cacao cultivated in the pacific region,18 aiming to enable cacao
its designation remains an ongoing task. Despite a few examples farmers to consistently produce fine flavored cacao beans. These
2
wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

Journal of The Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.
Cacao denomination of origin www.soci.org
are examples of cultivated cacao, but Colombia is home to many although we do not claim to devise a comprehensive analytical
unexplored materials of wild cacao species,20 with numerous wild methodology.
genotypes from the Amazon still waiting to be discovered and uti- The proposed bottom-up framework comprises two inter-
lized in determining future designations of origin. connected cycles (green and blue cycles in Fig. 2). The first step
To date, no comprehensive characterization has been con- is a regionalization mapping, which involves scaling from a
ducted to define or classify cocoa quality based on its genetic national to regional level using climate and soil indicators specific
diversity, regional or local climate, soil characteristics, landscape to each geographic area (green cycle). Once the region of interest
features, sensorial and biochemical indicators, and productivity is identified, an agronomical assessment is conducted to charac-
aspects. This holds true not only for Colombia, but also for other terize the cacao cropping system, including the evaluation of
cacao-producing countries in South America. Consequently, the pests and diseases present in the region. Moving forward, the
series of steps and variables that our framework contains repre- blue cycle focuses on quantifying product typicity variables of
sent a significant advancement in the field of cacao DOs and cacao at both the individual farm and clusters of farms levels. This
GIs, particularly for unexplored geographic areas within the trop- cycle involves conducting a genetic identity analysis, a detailed
ical Pacific region of Colombia. The present study demonstrates typification of cocoa quality, and a survey of microorganism diver-
the feasibility of establishing cacao DOs and GIs by employing a sity, all of which are crucial in defining the sensorial characteristics
multidisciplinary and robust scientific approach that lays the of chocolate. All these steps are interconnected and collectively
foundation for supporting cacao DOs through a series of indepen- constitute the fundaments and key variables in the proposed
dent steps. framework that are required to achieve a well-established denom-
ination of origin for cacao.
MATERIALS AND METHODS Geography

Regionalization of GIs
Study region and sampling zones
The Buenaventura region is located in the department of Valle del A biogeographical regionalization approach was employed to
Cauca, situated in the southwestern part of Colombia along the identify distinct cacao regions in Colombia, which were subse-
Pacific Ocean (Fig. 1). Buenaventura is part of the Biogeographical quently designated as GIs.11 To conduct the regionalization, a
Chocó region.21 The region is encompassed by tropical rainforests dataset consisting of 4974 cacao farm occurrences of cacao across
that experiences some of the highest precipitation levels on Colombia, along with macroclimate variables,23 and microclimate
earth.22 The exceptionally high rainfall is a result of the conver- variables,24 were utilized as inputs. The macroclimate WorldClim
gence of vast amounts of air rising from the ocean, combined with v.2 dataset23 comprises seven variables, including precipitation,
maximum temperature, minimum temperature, mean tempera-
the intricate and rugged topography of the area, as well as the
ture, relative humidity, wind speed and solar radiation, with a spa-
intense evapotranspiration occurring within the tropical rainfor-
tial resolution of 1 km2 and a soil depth of 0–5 cm. The
est. Within the Buenaventura region, three specific cacao sam-
microclimate analysis employed the SBIO dataset,24 which con-
pling zones were selected, each representing a distinct cacao
tains eleven soil temperature variables. The macroclimate analysis
cropping system.
enabled an understanding of the climate variations between dif-
The average climate conditions for each study zone are pre-
ferent geographic regions, whereas the microclimate analysis
sented in Table 1. Zone 1 (Cisneros) is situated closer to the foot-
focused on determining specific microclimate conditions based
hills of the western Andean mountain range at an elevation of
on soil properties.
515 m a.s.l. It experiences the highest amount of solar radiation
Mean climate and soil temperature values for each cacao farm
and notable temperature variations between the maximum and
occurrence were extracted from the WorldClim2 climate raster
minimum values during the day and night. Zone 2 (Bajo Calima),
and the SBIO soil temperature raster. Principal component analy-
located on the lowlands closer to the Pacific Ocean at approxi-
mately 40 m a.s.l, is characterized as the rainiest and most humid sis (PCA) was performed using the extracted climate data from
among all three zones. Lastly, zone 3 (Sabaletas) is positioned fur- WoldClim2 and SBIO rasters, employing R, version 3.6.3
ther south on the marine plains of the coastal areas, with an eleva- (R Foundation, Vienna, Austria) and the ‘vegan’ package. The
tion of 15 m a.s.l. This zone exhibits the highest average PCA axis 1 results were then mapped onto the climate and soil
maximum temperature and is the warmest among the three. temperature values using a geospatial application.25 This applica-
The soils found in the cacao plantations of the study region are tion facilitated the generation of climate and soil temperature
primarily Inceptisols and Entisols. These soils are characteristic of maps for all cacao producing regions in Colombia, based on the
the area and are highly acidic, with pH levels from 5.0 to 5.4. Addi- principal components. These results serve as GIs, reflecting spe-
tionally, they occasionally show significant limitations as a result cific geographical origins, and were utilized to identify the cacao
of aluminium toxicity. Lime application and fertilization practices GI for Buenaventura at the national level. Lastly, a geospatial
are infrequent. The dominant soil textures in the region are loam local-level PCA analysis was independently conducted for the
and clay loam, and the organic matter content varies, typically three sampling zones using the same methodology, aiming to
ranging from moderate to high levels. characterize the finer differences in local climate and soils
between the three zones. This local analysis can be considered
as a fine-scale identification of GIs.
Summary of the proposed framework
Our framework is not a mere variable-based methodology but Agronomic baseline
comprises a series of steps to establish cacao denominations of A technical diagnosis of cacao farmers was conducted using a par-
origin in tropical regions. Cocoa quality, including other unique ticipatory methodology, where farmers provided descriptive
traits, is vital for such denominations. Our proposal integrates information regarding their agronomic practices and farming sys-
these variables, contributing significantly to the concept, tems. In total, 12 cacao farms for the three zones were visited, and
3
J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

Figure 1. Map of Colombia with the location of the Buenaventura study region on the Pacific coast and the three sampling zones.
Table 1. Average climate conditions for the three sampling zones
Maximum Mean Minimum Vapor Wind

Diurnal temp Solar radiation temperature temperature temperature Rainfall pressure speed
Zones range (°C) (kJ m−2 day−1) (°C) (°C) (°C) (mm) (kPa) (m s−1)
1. Cisneros 8.74 16.310 29.06 24.70 20.32 2686 2.52 1.05

2. Bajo Calima 7.65 15.766 30.15 26.33 22.45 6262 2.88 1.31
3. Sabaletas 8.25 16.123 30.54 26.37 22.24 4169 2.86 1.21
Note: Underlined values are the maximum averages of the three zones for each climate variable.
interviews were conducted with 32 farmers. Some of these approach that involved close collaboration between farmers and
farmers were members of cacao grower associations active in researchers. In total, 11 plantations were selected, with three
the study area, including ACABC (association of farmers and cacao located in zone 1, three in zone 2 and five in zone 3. The sampling
growers of Bajo Calima), APCC (Cisneros growers association) and took place between the 6th and 10th of October 2018, during the
Sabaletas Community Council. The cacao plantations in the study peak cacao harvest period of the year. In each plantation, 10 trees
region contained genetic materials of cacao related to FQ-1, were randomly chosen by following a zigzag transect pattern that
FEAR5, F61, SUI-83, SUI-62, GS, TSH-565, ICS-95, CCN-51, FCH-8, aimed to cover the entire field. The path and distance of the tran-
FQ-1 and FEE-2. sect were approximately determined beforehand during an
inspection of the plantation, with guidance from the farmer.
Pests and diseases For each sampled tree, we conducted a comprehensive assess-
The infestation level of the cacao pest species and the incidence ment of cacao fruit damage. Specifically, we recorded the damage
of fruit diseases were determined through a participatory caused by different categories of pests, including:
4

Figure 2. Flow diagram of the proposed framework.
(1) cacao mirids (Hemiptera: Miridae), Fig. S1E). When available, insect pests associated with the dam-
(2) sucking insects (other than mirids), that included treehoppers aged pods were collected for taxonomic identification.
and thorn hoppers (Hemiptera: Membracidae) as well as scale The primary disease issues affecting cacao, including (1) the
insects (Hemiptera: Coccidae) and mealybugs (Hemiptera: frosty pod rot (FPR), caused by Moniliophthora roreri (Cif. & Par.);
Pseudococcidae), (2) the black-pod rot (BPR), caused by Phytophthora spp.; and
(3) pod borers (Lepidoptera), (3) witches' broom disease (WBD), caused by Moniliophthora per-
(4) geometer moths (Lepidoptera: Geometridae), and niciosa, were identified, although not all were surveyed in the pre-
(5) rodents and other mammalians. sent study.
To determine the causal agent responsible for each disease,
To identify cacao mirid damage, we looked for the characteristic plant tissue samples were collected from infected plants. Dis-
dark, round spots on the pod surface26 (see Supporting informa- eased pods were carefully removed, including healthy tissues
tion, Fig. S1A). Damage caused by other pod-sucking insects extending approximately 4 cm beyond the lesions. These samples
was differentiated from mirid damage by its black to brown scar- were then placed in plastic bags and transported to the laboratory
ring, irregular shape and often connected scars, indicating the for further analysis.
previous presence of clustered mealybug or membracid colonies In the laboratory, the samples underwent surface sterilization
(see Supporting information, Fig. S1B). These scars sometimes had using a 2% sodium hypochlorite solution for 15 min. Subse-
yellowish–green halos, which were not observed around mirid quently, they were treated with 70% ethanol for 1 min and rinsed
spots. To confirm this type of damage, we removed mealybug three times with sterile water. The plant tissue was dried and then
and membracid colonies from the pod surfaces to examine the sectioned into six segments measuring 0.5 cm each. These seg-
scarring caused by their feeding activity. Because of their similar ments were directly placed on Potato Dextrose Agar (PDA; Oxoid,
appearance, this type of damage was grouped into a single cate- Basingstoke, UK) supplemented with 0.1% Triton 100X (Sigma-
gory and could not be distinguished without the presence of the Aldrich, St Louis, MO, USA) and 0.3% chloramphenicol to prevent
hemipteran colony. bacterial contamination. The plates were incubated at 25 ± 2 °C
Pod borer damage was characterized by black entry spots under 24 h of light for 7 days. Any fungal growth observed was
accompanied by humid frass buildup in their vicinity, often with sub-cultured onto fresh PDA or water agar (1% w/v) until pure col-
multiple spots per pod. We checked these spots for tunneling onies were obtained.
activity and the presence of larvae. Geometrid damage (see Sup- To compare infestation levels across the different study zones,
porting information, Fig. S1C) was identified by cherelles or pods we employed generalized mixed effect models (GLMM) with a
with gnawed scars, resulting in wide depressions or holes on the Poisson error distribution and a log-link, utilizing the ‘lmer4’
smooth surface (the latter likely caused by late instar larvae). Dam- package. The experimental unit was defined as the individual
age to cherelles and young fruits often led to misshapen fruits as tree, with the number of pods displaying signs of pest and dis-
they matured (see Supporting information, Fig. S1C). Geometrids ease damage compared to the total number of pods per tree.
were observed on leaves, fruits and leaf surfaces, where they con- This ratio was used as an offset variable.27 The sampling zone
sumed the plant material. was included as an explanatory variable, and the plantation
Finally, rodents created large holes in the pods to access and was treated as a normally distributed random effect to account
consume the pulp and seeds (see Supporting information, for any spatial autocorrelation among trees within the same
5

transect. Pairwise post-hoc comparisons between zones were The fermentation and drying processes were conducted as fol-
conducted using Tukey's tests on the model estimations. Models lows: Cacao pods were carefully opened using sharp knives, and
were assessed for overdispersion, and the Pearson residuals ver- the seeds were manually extracted and mixed to create an initial
sus the fitted values were visually inspected to validate the mass comprising all the collected cacao cultivars. From this
models. initial mass, 50 kg of seeds were extracted and placed into six sep-
arate compartments within a wooden fermentation box (Fig. 3),
Genetics allowing for six repetitions of the process. The fermentation fol-
Plant material and DNA extraction lowed a standard method widely employed by cacao producers
Leaf samples were collected from the three study zones to obtain in Colombia.39 Initially, the cacao mass underwent an anaerobic
genotypic information on the cultivated cacao trees in the region. stage lasting 48 h without any mixing. Subsequently, mixing
Leaves were sampled from ten farms in zone 1, six farms in zone and aeration of the cacao seeds were performed every 12 h for
2 and 10 farms in zone 3. DNA extraction from the leaves and subse- a total duration of 120 h. Throughout the fermentation process,
quent genotyping was performed via LGC Biosearch Technologies the temperature of the fermentation mass was monitored at the
(Teddington, UK). KASP genotyping assays were utilized to identify central point of each compartment using a thermocouple (model
single nucleotide polymorphisms (SNPs) at specific loci of the cacao 735-1; Titisee-Neustadt, Germany). Finally, the cacao samples
genome, which had been previously used in various studies.28–32 Bi- intended for sensory and metabolomics analysis were dried in a
allelic discrimination was achieved by competitively binding two forced convection oven at 50 °C, utilizing hot air.
allele-specific forward primers, each having a unique tail sequence During the sampling process, 300 g of cacao beans were
that corresponded with two universal fluorescence resonant energy extracted from the central portion of the cacao seeds mass in each
transfer cassettes. One cassette was labeled with FAM dye, whereas of the six compartments of the fermentation box. These beans
the other was labeled with HEX dye. were specifically designated for sensory analyses. Additionally,
50 g of cacao seeds were extracted for untargeted metabolomics
Phylogenetic analysis analyses.
Out of the initial 96-SNP panel selected for studying genetic diversity,
87 SNP markers were utilized to align the new dataset generated in Quality analysis
the present study with the dataset published previously.31 The align- For the descriptive sensory analysis and evaluation of chocolate
ment of the SNP dataset was performed using ClustalX, version flavour profiles, chocolate bars containing 67% (w/w) cacao solids
1.83,33 and visually confirmed using Geneious, version 8.1.7.34 After were produced. The laboratory of sensory analysis at CIRAD
concatenating the data, a phylogenetic tree was reconstructed using (Centre de Coopération Internationale en Recherche Agronomi-
the maximum likelihood method with PhyML, version 3.0,35 imple- que pour le Développement, Paris, France) employed a standard-
mented in a bioinformatics platform (http://www.atgc-montpellier. ized protocol.18 The descriptive sensory analysis was conducted
fr/phyml). The phylogenetic analysis involved 100 bootstrap repli- using the consensus profile method outlined by the International
cates, and gap sites were treated as missing data. The species Theo- Organization for Standardization (ISO).40
broma grandiflorum (Wild Ex. Spreng. Schum) was used as an The blind tasting sessions involved 8 trained judges who evalu-
outgroup in the phylogenetic tree. ated the chocolate samples in duplicate. Two evaluation sessions
To assess possible significant genetic differences between the study took place. In the initial session, the judges agreed upon the sen-
zones, a one-way permutational multivariate analysis of variance sory attributes to be assessed. Subsequently, each judge individu-
(PERMANOVA) was conducted, and the data was visualized through ally evaluated all the chocolate samples, assigning scores on a
PCA using Past, version 4.1.36 The PCA analysis was based on the scale from 0 to 10 for the predetermined sensory attributes. Cron-
sum of SNP differences between samples. Furthermore, the ancestry bach's alpha coefficient was used to assess the reproducibility of
of each genotype was analyzed using the intersected SNPs represent- the judges’ evaluations.
ing the 10 recognized cacao genetic groups.37 A maximum likelihood To confirm the assumptions of normality and homogeneity of
analysis was performed using Admixture, version 1.3,38 in a supervised variance, Shapiro–Wilk and Levene's tests were conducted.
mode, where the reference genetic groups were specified. Because these assumptions were not met, a non-parametric test,
specifically the Kruskal–Wallis test, was employed to determine
Quality significant differences among the chocolate samples from the
Typification various geographic zones. A Bonferroni correction was applied
One post-harvest transformation site for each zone was con- to post-hoc test pairwise comparison zones.
ducted to evaluate the quality potential of cacao cultivated.
Figure 3 illustrates the stages involved in the process. The assess- Untargeted metabolomic analysis: the chemical fingerprint of
ment of quality potential involved studying the flavour profiles of fermented and dried cacao beans
chocolate (or cocoa) and analyzing the cocoa metabolome after Extraction, instrumental analysis, and data processing procedures
post-harvest processing. The entire process was carried out at a followed stablished methodologies.19,41 Initially, all samples were
single site in each region, utilizing the same fermentation meth- ground and defatted. An extraction solution composed of ace-
odology and equipment. tone water:formic acid (70:29.5:0.5) was employed, and ultrasoni-
cation was utilized during the extraction process.
Postharvest transformation process For instrumental analysis, an ultra-performance liquid
Harvesting involved selecting a representative sample of cacao chromatography-electrospray ionization (UPLC-ESI)/quadrupole-
cultivars from each of the 32 farms in each study zone. Healthy time-of-flight-mass spectrometry (MS) system (Acquity; Waters
and fully ripe cacao pods were carefully harvested one day prior Inc., Milford, MA, USA) was employed. Chromatographic separa-
to the fermentation process and then transported to the central tion was achieved using a UPLC CSH C18 column (1.7 μm particle
postharvest transformation site. size, 2.1 mm × 100 mm; Waters Inc.) coupled with a pre-column
6

Figure 3. Scheme of the methodology developed to analyze the quality potential of the cacao cultivated in the selected farms from Sabaletas, Cisneros
and Bajo Calima.
VanGuard CSH C18 (1.7 μm particle size, 2.1 mm × 5 mm; Waters 0.85 g per 100 mL of NaCl in 250-mL Erlenmeyer flasks using a
Inc.). A gradient elution of water and acetonitrile (solvent B) was rotatory shaker (Thermo Scientific MaxQ 4000; Invitrogen, Burling-
utilized. ESI-MS analyses were conducted in positive ionization ton, ON, Canada) at 150 rpm for 30 min. The resulting fluid was
mode, employing a full scan range of 100 to 1500 Da in continu- decanted and filtered through sterile gauze. The free-pulp solu-
ous mode. To ensure quality control, two pooled samples were tion was then centrifuged at 3220 × g at 12°C for 20 min
measured after every 10 samples. (Hermle, Z32HK; Labortechnik GmbH, Wehingen, Germany) and
Data processing involved the use MassLynx, version 3.1 SCN the pellet obtained was resuspended in 10 mL of sterile saline
639 (Waters Inc.) and Progenesis QI software (Waters Inc.) to select solution. DNA extraction from this solution was performed using
and align chromatographic peaks. The effect of batch-related drift an UltraClean Microbial DNA isolation kit (MoBio, Carlsbad, CA,
on the quality control baseline was corrected using Metabodrift.42 USA). The concentration and purity of the extracted DNA were
Data normalization was performed using the LOESS (i.e. locally quantified using a NanoDrop 1000 Spectrophotometer (Thermo
reweighted scatterplot smoothing) method (⊍ = 0.5). Finally, the Fisher Scientific, DE, USA) and stored at −20 °C.
data were transformed using a dilution factor that accounts for The bacterial communities associated with the cacao bean mass
the mass of the cacao powder and the mass of the sample upon were analyzed by sequencing the hypervariable V4 region of the
resuspension. 16S rRNA gene. Bacterial libraries were prepared following
To integrate the extensive metabolomic data and conduct mul- the protocol previously described44 which involved a two-step
tivariate statistical analysis, a PCA was employed to visualize the PCR procedure. In the first PCR, modified primers 515F-806R,45
data patterns and trends. Subsequently, a PERMANOVA was con- with a linker region in the 50 end were used. In the second PCR,
ducted to identify significant differences in the metabolomes of barcodes, as well as the Illumina i5 and i7 regions, were added
the fermented and dried beans between the three zones. Addi- to the amplicons and followed a specific protocol.46 The libraries
tionally, partial least square discriminant analysis (PLS-DA) was were sequenced using an Illumina MiSeq with the kit V2
utilized to evaluate the relationship between the metabolomes (Illumina, San Diego, CA, USA) (2 × 250 bp).
and the different zones, as well as to assess the predictive ability Bioinformatic analysis was conducted to taxonomically identify
of the classification model. the composition of the bacterial communities. First, the samples
were demultiplexed using an in-house script. The quality of the
Microorganisms (bacteria) sequences was assessed using FastQC v0.11.7 (http://www.
The bacterial cells associated with the cacao bean mass were sep- bioinformatics.babraham.ac.uk/projects/fastqc). Adapters and
arated from the samples using the protocol previously linker sequences were removed from the partial sequences of
described.43 Briefly, frozen beans (20 g) were homogenized with the 16S rRNA gene using Cutadapt, version 1.12.47 Low-quality
7

sequences, with an average Phred score < 20 and a read Agronomic baseline
length < 200 nucleotides, were identified and removed from The types of cacaos found in the whole study region were FQ-1,
the datasets using Trimmomatic, version 0.38.48 The Qiime 2 FEAR5 (Trinitario), F61, SUI-83, SUI-62, GS, EET-72, TSH-565
pipeline v_2018 (https://forum.qiime2.org/t/qiime-2-2018-8- (Trinitario), ICS-95, CCN-51, FCH-8, FQ-1 and FEE-2. In zones
release-is-now-live/5860) was employed for the analysis of the 3 and 2, the main cacao cultivars grown were ICS 95 (Trinitario x
16S rRNA gene amplicon libraries. The Dada2 (https://qiime2.org) Criollo) and CCN 51, which were planted in the context of a rural
algorithm was used to denoize the sequences and generate development project aimed at encouraging cacao cultivation.
amplicon sequence variants (ASVs). The pair-end sequences were Most farms in the region were small-scale, ranging from 1 to
inputted into Dada2 with a truncation length of 200 bp. ASVs that 4 ha in size. The average age of the cacao plantations was approx-
were present in fewer than two samples or had less than four imately 9 years, but most farmers did not maintain records of the
sequences in all the samples in the feature table were excluded exact number of cacao trees on their farms. Excessive shading was
from further analyses. The taxonomic assignment of the ASVs observed in certain farms in zones 2 and 3, likely as a result of the
was performed using the Silva Database, version 132.49 efforts to mitigate the effects of high rainfall. This shading can
affect cacao production. Diseases such as witch's broom, frosty
pod rot and black-pod rot also require intensive agronomic man-
agement to mitigate their impact on the plantations. The three
RESULTS AND DISCUSSION
study zones featured diversified farming systems, with a focus
Regionalization of GIs
not only on cash crops, but also on other native crops for domes-
The GI for the Buenaventura study region was identified as part of tic consumption. The most common were peach palm (Bactris
the cacao southwest region, specifically the 4C sub-region and gasipaes Kunth), coconut (Cocos nucifera L.), cassava (Manihot
central Calima province number 43 in the department of Valle esculenta Crantz), citrus fruits (Citrus spp.), borojó (Borojoa patinoi
del Cauca.11 The climate and soil temperature conditions in this Cuatrecasas) and Chinese potato (Coleus parviflorus L.). Some
province are characteristic of the lowland rainforests in the Pacific farmers also mentioned the presence of timber trees such as
region. However, the GI for zones 2 and 3 differs from that guino (Carapa guianensis Aubl.) and sajo (Campnosperma pana-
observed for zone 1 as a result of their lower elevation. Soil tem- mense Standl.).
perature seasonality is lower in zones 2 and 3 compared to zone 1.
The local level geospatial PCA analysis demonstrated that each Pests and diseases
sampling zone possesses specific climate characteristics (Fig. 4). In total, 1358 cacao pods were assessed for pest and disease dam-
Consequently, the study zones have distinct climate GIs, as indi- age. Regarding pests, rodent-damaged pods were only recorded
cated in Table 1, which are further supported by the national level in a single plantation in zone 2, with two pods (1.5% damage)
regionalization. Solar radiation and daily temperature range so they were not statistically analyzed. In the overall study region
emerge as the primary driving variables in zone 1, suggesting of Buenaventura, geometrid and sucking insect damage (exclud-
fewer extreme conditions compared to the other zones. Con- ing cacao mirids) were the most commonly observed on cacao
versely, zone 2 is strongly influenced by high levels of rainfall pods, accounting for 19.2 ± 1.9% and 17.9 ± 1.9% of the pods,
and humidity. Temperature exhibits the most significant effect respectively (mean ± SE). Pod borers and signs of cacao mirid
on zone 3. feeding activity were much less frequent, present in only 3.5
± 0.8% and 3.4 ± 0.3% of the pods, respectively.
Zone 1 had higher infestation levels of cacao mirids and pod
borers compared to the other two zones (Fig. 5A,D). No pods dam-
aged by these pests, nor the insects themselves, were observed in
zone 2. However, zone 2 recorded the highest levels of fruits dam-
aged by geometrids compared to the other zones (Fig. 5B). In
these plantations, a large number of geometrid larvae were
observed feeding on cacao foliage (M. Porcel, personal observa-
tion). Zone 2 also had a high number of pods with signs of sucking
activity, presumably from membracids. Zone 3 showed high activ-
ity of geometrids and membracids (Fig. 5B,C), whereas a moder-
ate number of pods damaged by cacao mirids and pod borers
were observed (Fig. 5A,D).
Pest damage was similar between zones 2 and 3 but differed sig-
nificantly from zone 1, which showed higher damage caused by
Monalonion spp. and cacao pod borers, which are well-established
pest species in the country.50 These differences in pest damage
could be attributed to the climatic conditions determined by altitude
(Fig. 4). Previous studies on cacao pests in other regions of the coun-
try have primarily focused on inland cacao plantations that have cli-
matic conditions more similar to zone 1. Therefore, the high levels of
geometrid and sucking insect damage (mainly caused by the mem-
bracid Horiola picta Coquebert) could be associated with the high
Figure 4. Local level climatic zones determined by a geospatial analysis
humidity and temperatures found in these zones. Under such condi-
of climate mapped as a principal component analysis of axis 1 (% explained tions, there may be a higher turnover of cacao damage, particularly
variance by the principal component). in the case of geometrids, which can cause severe deformations to
8

Figure 5. Proportion of cacao pods (mean ± SE) with feeding damage caused by (A) cacao mirids, (B) geometrid larvae, (C) sucking insects (other than
mirids) and (D) lepidopteran pod borers and incidence per pod of (E) frosty pod rot disease, (F) black pod rot disease and (G) witches' broom disease.
Different letters indicate statistically significant differences (GLMM: Tukey's test, P < 0.100). No letter is assigned when the percentage is zero.
pod and consume entire cherelles. The cacao mirid species identified Previous reports have already highlighted that FPR disease is one
in the study region was Monalonion dissimulatum Dist, which is the of the main factors limiting yields in tropical America.52-53 It is twice
most common species associated with cacao in the neotropics.51 as destructive as BPR and more challenging and dangerous to con-
All the pod borers identified belonged to the Ecdytolopha genus, trol compared to WBD. This poses a serious challenge for cacao
whereas species of the Carmenta genus were not found in the pre- growers in the region, emphasizing the urgency to develop new
sent study. cacao materials that are resistant to FPR.
As part of the participatory research approach, it was observed that,
although farmers were able to identify unhealthy cacao pods, they Genetics
struggled to determine the specific type of pest or disease affecting Phylogenetic analysis
the plants. There was significant variation in the way farmers managed The maximum likelihood phylogenetic analysis resulted in a tree
unhealthy plants in the field. Some farmers mentioned following man- with low bootstrap values, indicating insufficient support for the tree
agement methods recommended by agrochemical dealers, whereas nodes. The samples from Buenaventura were distributed across dif-
others relied on traditional approaches for disease management such ferent branches, suggesting that they likely have diverse genetic
as sanitary harvesting. Additionally, important agronomic practices backgrounds and belong to distinct cacao genetic groups (Fig. 6).
such as pruning and removing cacao pods affected by pathogens Farms in zone 1 exhibited materials related to those from the
were not commonly practiced in the region, leading to a continuous Colombian Cacao Federation (Fedecacao), including FQ-1 and
presence of diseased pods and facilitating the spread of pathogens. FEAR5, as well as Colombian hybrid commercial materials such as
Cacao growers expressed that the implementation of more rigorous F61, SUI-83, SUI-62 and GS. One tree in zone 1 was potentially asso-
agronomic practices increased production costs, thereby reducing ciated with Ecuadorian cacao EET-72. Materials from zone 2 showed
profitability. However, a few cases of farmers with better management close relationships with Ecuadorian materials (EET), the Trinidad
practices were also observed. commercial material TSH-565 and materials from Fedecacao. In zone
FPR was identified as the most significant disease affecting the 3, the materials were related to ICS-95 and CCN-51. Other cultivated
cacao production system (Fig. 5E). In mature pods (2.5– materials demonstrated stronger connections with Fedecacao, such
3.5 months old), brown spots with irregular edges were observed. as FCH-8, FQ-1, and FEE-2, whereas one material was linked to cacaos
The severity of the disease varied, but many infected pods were in from Tumaco (Colombian Pacific coast region). The PCA results
the sporulation phase, which is critical for the optimal spread of (Fig. 7) indicated that the samples did not form distinct clusters
spores by the pathogen. This finding indicates that infected pods based on their zones, and the PERMANOVA test confirmed no signif-
attached to the trees were not effectively removed. In total, 33 iso- icant genetic differentiation between the zones (P = 0.19).
lates were obtained from three study zones. Macroscopic charac- In the supervised mode of population structure analysis, we
teristics confirmed that the isolates collected corresponded to were able to determine the ancestry of the samples from Buena-
Moniliophthora spp. for FPR and WBD. However, it was not possi- ventura (Fig. 7). Our findings suggest that the materials found in
ble to isolate Phytophthora spp. from cacao pods affected by BPR. this region are likely the product of admixture. All the genotypes
9

CHOCOLATE_4
CRICF_37
CRICF_30
CRICF_11
CRICF_26
CRICF_46
CRICF_31
CRICF_50
CRICF_8
CRICF_4
MUZO_6
CRICF_16
CRICF_ 14
SCC_20
CRICF_ 48
CRICF_1
CRICF_ 1
CR_35 56
861648
830758
SCC_5
SCC_6
861646
871868
UF_676
CRICF _44
830759
SP_M2_
CRICF _51
CRICF _28
UF_677
CRICF _32
UF_667
UF_650
CRICF _33
C_100
CRIC _35
R_30
CRIC F_36
R_41
R_13
UF_12
CRIC F_43
CRICF
CRIC F_49
8307632
R_117
EET_ _9
CRIC F_55
R_105
FCM_ 9
CRIC F_12
CRIC F_18
EET_ 8
EET_ 2
R_2
CRIC F_20
CR F_45
CR ICF_10
SC_2 7
CR ICF_27
SC 1
ICA_ 6
SC_2 5
CR ICF_34
SC_2
33
CR ICF_40
SC_1 8
CR ICF_41
SCC1371
SC_140
8613_77
CR ICF_42
EET_
CR ICF_52
EET_ 8_1
86 253
SP_M
CR ICF_958
SP_MS_6 9
CR ICF_
7
CR ICF_5 4
CR ICF_7
IC _4
CR _38 7
CR ICF_5
352
SC _46
CR ICF__25
SC _47
8_20
SC _45
CR ICF _53
SC _43
CR F 6
CR ICF _15
SC _42
CR ICF
SC _4
C ICF _47
C RICF _24
SC _33
SC 34
C R_2 _19
IC
SC C_3
GO
_
SC C_5
S _35
C R_1 DIE
S C _ 29 2
S _44
CRR_11
SC _
C ON_ 3
D _4
F CM F_5
S SV 15 5_1
SC_16
SCC_156
TUCC__155 5_2
S A_ 5
F SV_ _23
06
S M 6 6 _3
IC A_ 4
F RIC
F SV_ 15
C CC CO 8 N_5
IC A_ 3
CHCM ETR O_7

IC A_ 2
_2
IC A_ 8
S O _1 A
IC 71 F 6 ATE
IC _4 6
F O P AC
SC C_
8 RIC _4 L
S
SU _ 9 2
CR CM C_ 11
IC S _3 4 3
C F_ 6
IC S 76 _1
F S _ 3
I_2 9
IC _2 7
IC C_ _53 0
IC S _7 9
A 5
S C 3
S A _4 8
SC SC 26 3
S C_ _ 1 0
A
EESPA C_ 40 7
IC C_ 10 0
IL
74
T_ _1 50
U
_ 9
_H
CH HO _12 2
U L O 7
S C_ 3 1
C A 1
H OB OC O_
A ION
S C_ _3 2
C
S C _2 8
2
A_ C
C 3_ 3
S C _3
6
IL A
C
R
15 53_ _1 1_
S F
R
IC IC
a_
SC F_ V_ _1 53 erl
P
3
SC SC _19 9 FS SV V_1 5 _P
_ F S 31 6 La
SP FCICA _20 1 F 61 40 1 2 a_
8 61 61 8_ lim
F C _ _1
_ M _1
FY SPA M6 _8
FY YC_ _2_ 19
8 61 _ Ca
8 ET _ _1
E ajo _8 _3 _2
B E T _8 0 4 _ 3
A
FC C_ 2_ 1 E ET _1 04 _1
2 2 E VA _1 04 _1
8 M _ F VA _1 14 2
SC 61 _30 3 F YA C_ 2_
C 62 F MA C_ 2 _3
P L S EE E _ 3 6 8 F MA _3_ 14
AY P T ET F LE C_
A_ _M _2 F MA 48
A 5 4 F S_ 8
86 LTA _1
PU 8 16 IC C_ 8
EB RIO 86 616 31 S A_ 7
LO _G 13 36 IC C_2 3
_B UA 19 S C_1 9 2_
5
5
EL P S C_3 72 1 ir_ _5 3_
AP UF ICS_ O
L I S 13 1_ en ir_5 o_
CA APO _60 8 86 P_M 4 _Porvorvensfuerz
TO RIS 2 S C_2 leta _P L_E
N S aba letas _M
MA SC GO S ba tas
R _5 Sa bale 5_A
84 TINA 5 Sa S_9 5
0 7 9
ICAGS_178 IC S_ 8
IC S_9 4
_0 2
C_ 7 0 IC 161 _2
C_ 58 86 _M1
9 SP A_13
UF C_967 IC I_62 la_4
_ SU one ea_1
GS 700
GS _10 Citr onel a_3
SC _06 Citr onel
SC C_8 Citr I_83
Cisne SC C_65 SU I_44 1
C_6 SU _M9_
ros_ PER GS_3 0
El_A SP 1365
guac PERSSEA_26 86 40 1
atic EA_1
o_1_ 8307 10 ALK_
G 2 8307 TA_W
EET_S_60 LOLIM_19
F
353 FC C_21
TRINCC_18 SC _M14_1
ID SP C_69
8616AD SC
19 C_ 23
CHUA861309 SC 8
O_
TSH_ 120 SC_3 70
66 SCC_ _M3_1
ICS_ 0 SNSM2 _2
PA_3 6
P SC_344 atico_2
8616 5 guac
45 8718 ros_El_A
8613
17 Cisne _M3_2
861613 SNSM 33
861623 FCM_ 5
Bajo_C EE SUI_7 45
alima_L EE T_6 SC C_
a_Perla T_395
_2_6 83076174
EET_67 EET_3
EET_19 861326 _3
3 TUMACO
ICS_1
F_6 SCC_83 6
861312 TSH_51
SCC_44 FCM_20 4_1
SNSM_M 1
8718
SNSM_M5_63
SP_M5_2
2
SNSM_M5_
ICA_219
B
CC_10 TSH_622
PENTAGONA_
VERDE EET_387
UF_273 SCC_59
SC_41 F_55_2
F_53 F_55_1
861310 FCM_40
F_6_4
FLE_3 TSH_565
861321 TSH_565_1
EBC_138 TSH_565_2
EBC_91 TSH_565_3
EBC_148 Bajo_Calima_La_Perla_5_6
EBC_143 FCM_39
EBC_144 FMAC_11_3
EBC_147 FMAC_11_2
EBC_121 871855
55
EBC_279 861311
EBC_101 830721
EBC_77 TUMACO_5
EBC_76 830724
EBC_06 C_95
EBC_48 TUMACO
EBC_29 830702
EBC_09 830715
TUMACO
830698 _2
TSH_81
861345 2
830729
830704
EET_7
Citron 5
EET_ ela_2_Fabio
UF_7 12 _Cong
a
EET_ 22
EET_ 338
EET_ 174
EET_ 30
UF_2 162
12 UF_6 21
SCA_A_6 EET_ 13
SC 64 EET_ 228
F_
F_582 EE 184
39 EE T_182
EET_ 716 FC T_313
NA_0713 SM M_1
83 36 CG _1
8307397 EET_ _12
EET_0720 F_18 17
83 4_2 F_11 0
_M EE 0
SM 830711 EE T_58
SN TIN
AR 54 EE T_61
N_M _6 EE T_3
SA TSAA_212 EE _4T 30
088
IC F_2947 EE T_9 8
00 EE T_625
C_ _2
AB _M7 _1 EE T_5
SP _M7 12 Ba T_9 9
SP CC_ _1 EE jo_C 6
S 13 1 EE T_9 alim
_M 14 EE T_9 6_1 a_L
SP SA_ 641 a_
T A_ 64 EE T_ 6_3 Pe
TS F_1302 EE T_ 96_2 rla
F_ 399 83 T_5238 _6
_6
0
C 69 3
T_ 00
EE T_4 777 E AP_ 9
EE 840 _18 83 BC_ 34
A 8 83 07 27
SP PA_ 19 Sa 070 16
C
S I_ 6 P ba 3
SU _25_2 TUOBL leta
T S M AC M _
EE ELA 171 7 8 P_ AC IO L_
ZU T_ P_ 9 8 307 M6 O_ N_ Esfu
N E EE 1
T_ 13 8 616 46 _2 6 TUM erz
VE EE P_ 272 T 616 41 AC o_2
9 S SH 12 O _5
T_ A_ 4 S CC _9
EE SC _3 86 L CC _5 42
Y _ 9 8 UK _7 8
NA F I_3 4 86 616 ER 6
NA _
SU PA 14 1 S 1 3 _5
S A_ _ SC CC 622 2 0
S C _
SP ON A_9 3 S UI_ _ 64
M P A_ 2
S P 2 SP P_ 72 79
P M
S A_ 20 7
_ T A_ _M 3_
SP PA _1 16 S S 6 3_ 2
S PA _ 11 Ba C_ H_7 7 1
S PA _ jo 51 92
_C
FC PA A_ 10
S PA
FL jo _3 30
86 M_ _15 2
S SP _
Ba LE C_1 3 _2
S ali
E_ _C _3
A
F B _ 41 _7
m
SP
3_ ali
EB SC 161 29
a_
E M _ 1 3 L_
83 C_ _11 8
1 ma
FC SV _4 1_ _M
La
F SV _4 ta
_P
IM 307 747 0
F V le
8 0 7
er
FS aba 8_ 2
_L
SU SC _6 5
la
S H 8_
4
a_
7
_4
FC CH 8_3 1_1 2
S I_1 _54
Pe
_6
F CH _5 1_ 3 orve enir _3_
S UI_ 00
C
F CN _5 1_ _P rv nir
rla
SUUI_ 24
C CN _5 tas _Po rve

I_ 69
_3
C CN ale s Po
1
S F_ 33
C ab leta s_
_6
T S UI_ 55
S aba leta1
F SH_ UI_ 18
S aba _ 5
Es
FSSA_ 118 27
S CN 54 s_M
FS A_ 13 8
fu
C 16 ro
FS A_ 13_ _3
86 isne S _1
erz
FS A_1 13_ 2
C AR _5 2
A 2 1
FEEAR _5_ 2
o_
SC _12 _1
F AR 2_ 3
FS C_ _3
4_
FE A_1 _5_
FS A_ 18
FS AR 89
5
P_ A_111
FE C_ 3
16 3
SC C_1 10_1 La_P

P _A
nir _4_ 5
SC _M ros_
SC C_112
SPisne _1
SC C_3 4
_1 5
on
C 1_4 _2
EB C_2 2
S _
F6 1_4 _3
_5
SC C_5 4
F6 1_4 9
te_
D
SC C_1 8
F6 C_1 s_2_
SC C_8 5
SC nero 1
C 5
De
Cis C_6 11_1
AR CAU _86
SC AC_ _40
9
FM R
_S
LU _7
GS A_1
AU _3
PA _29
RO 6
PA _169
SC M_22
ion
EB _121
FC A_2
K
19
FT
C
PA 377
E
FT
_1
_5
T _2_1
TSA _150
FT
TA _2_3
C_1
_644
FT
lay
TO 18
TA 2_2
_1
FCM leta_M
UF_ _5
TA_ _15
Saba 2_1
EET_
P_39 29
FEE__12
PA_4_A
a_
P_
FCM_
2_A
RO_5_38
PA 6
FCM 9
PA_1_81
SUI_ 9
8616 34
_1
SUI_983
FCM_BA
6
S_OI 28
SCC_ 75
FCM_ 5
nib
SU 47
FCM_14
_6
SUI_9 I_99
8307 1_1
SU A
FMT_ 1_3
SP_M I_61
FMT_ 1_2
SCC_
FCM_ _1
FMT_
CEPE 35
al_
861370
6_
SP_M11_46
EBC_125
8718541
SCC
SPA
FCM_13
SOPETR
FCM_31
SPA
12
SCC_91
SPA
830752
TSA_634
830755
P _5 AN_12
Mo
SPA_
C_
SCC_74
FQ_1_1
FQ_1_2
FCM_17
P _6
FQ_1_3
ARAUCA_3
830723
L_Es
Cisneros_1_1_Aniba
P _12
Sabaleta_ML_E
FMAC_12_2
ABC_062
ABC
C_4
FMAC_12_3
AMELONADO
861610
FMAC_12_1
SP_M4_2
SNSM_M1_1
SP_M4_1
SNSM_M1_2
FCM_24
P 7
SNSM_M2_2
SNSM_M2_1
FCM_11
830706
rale
411
fuer
s
062
zo_5
_5
sfuerzo_1_5
l_Morales
Figure 6. Maximum likelihood phylogenetic tree showing the relationships of the samples from Buenaventura (green, red and blue labels) with
AGROSAVIA cacao germplasm.
exhibit mixed ancestry from the Amelonado and Criollo genetic study, an integrated approach combining untargeted metabolomics
groups, with the majority also displaying ancestry from the Iquitos and multivariate statistics was employed to analyze the chemical
and Contamana genetic groups. Only two materials, one from profiles that define the unique characteristics of cocoa products from
zone 2 and one from zone 3, showed apparent ancestry different agroecological regions in Buenaventura.
from the Nacional genetic group. These results indicate that the After processing the raw files, approximately 4400 features were
cultivated genotypes in these farms have a hybrid origin, combin- detected within the mass range of 100–1500 Da in positive mode.
ing genetic traits from different cacao genetic groups. The unsupervised PCA revealed distinct clustering of metabolo-
mic fingerprints among the regions (Fig. 8A). The majority of the
Quality analysis variance was explained by the first two components (Fig. 8B), with
Untargeted metabolomic analysis to define a chemical notable differentiation observed between zone 1 and zones 2 and
fingerprint of cacao 3, as confirmed by the significant differences identified through
Metabolomics serves as a valuable tool for evaluating the impact of the PERMANOVA analysis (P < 0.001). Each metabolome repre-
various factors, including geographical origin and postharvest pro- sents a unique chemical fingerprint of cacao from this specific
10
cessing stages, on the biochemistry of cacao beans. In the present area of Colombia, as validated by the adjusted PLS-DA model.

1.00
Proportion of ancestry
Genetic groups
Amelonado
0.75
Criollo
Purus
0.50 Iquitos
Nacional
0.25 Parinari
Guiana
0.00 Nanay
Bajo_Calima_1
Cisneros_1
Sabaleta_ML1
Sabaleta_P1
Amel_1
Criollo_1
GU-1
IMC-1
LCTEEN_1
NA_1
Nacional_1
PA_1
Purus_1
SCA_1
Sabaleta_M
Bajo_Calima_2
Cisneros_2
Sabaleta_P2
Amel_2
Criollo_2
GU-2
IMC-2
LCTEEN_2
NA_2
Nacional_2
PA_2
Purus_2
SCA_2
Bajo_Calima_4
Cisneros_4
Sabaleta_ML2
Amel_4
Nacional_4
PA_4
Cisneros_6
Sabaleta_ML4
Sabaleta_P4
Criollo_4
GU-4
IMC-4
LCTEEN_4
NA_4
Purus_4
SCA_4
Bajo_Calima_3
Bajo_Calima_6
Cisneros_3
Amel_3
Nacional_3
PA_3
Purus_3
SCA_3
Sabaleta_ML3
Sabaleta_P3
Criollo_3
GU-3
IMC-3
LCTEEN_3
NA_3
Bajo_Calima_5
Cisneros_5
Amel_5
Criollo_5
GU-5
IMC-5
LCTEEN_5
NA_5
Nacional_5
PA_5
Purus_5
SCA_5
Sabaleta_Pr
Contamana
Curaray
Buenaventura's materials
Figure 7. PCA of the population structure of the Theobrom cacao materials using single nucleotide polymorphisms (SNPs). Colors in each bar correspond
to the probability of belonging to a reference genetic group. The pure clusters on the right side correspond to the reference samples.
Figure 8. (A) PCA score plot of the metabolomic fingerprints of the fer-
mented and dried cacao samples from each zone; (B) Scree plot of the var-
iance percentage explained by each component. The first two PCs
explained 45.4% of the total variance.
The model demonstrated high predictive power, with a cross-

validation using the leave one out algorithm (LOOCV) yielding a
predicted residual sum of squares (Q2) of 0.884. Moreover, the
individual component R2 indicated robust model fitting, explain-
ing 0.962 of the variance.
These chemical fingerprints consist of low molecular weight
metabolites known as chemical flavor precursors, which play a Figure 9. Dynamics of temperature (red line) and pH (blue line) during
crucial role in the non-enzymatic browning reactions during sub- the cacao fermentation process in the three zones: (A) zone 1, (B) zone
sequent stages, such as roasting (see Supporting information, 2 and (C) zone 3.
Fig. S2). Optimal conditions of temperature, time and water activ-
ity facilitate the interaction of these metabolites in Maillard reac- associated with distinct sensory descriptors of cacao, including
fine, bulk or undesirable flavors. Additionally, non-volatile
11
tions, leading to the generation of volatile aroma compounds

(A) (B) (C)
Figure 10. Bacterial succession during cacao bean fermentation. The relative composition of the bacterial communities throughout the cacao fermen-
tation process is shown for (A) zone 1, (B) zone 2 and (C) zone.
Acidity
Global quality 7 Vegetal
6 Sensory Bajo
Cisneros Sabaletas
5 attribute calima
Spices Grilled
4 Fine flavor
3 Fresh Fruits ab ab c
Florality 2
Animal
Nuts a a a
1 Floral a a a
0 Spices a a a
Nuttiness
Global quality ab ab c
Bitterness
Basic flavor
Sweet a a a
Astringency Sour a a a
Fruitiness
Bitter a a a
Cocoa taste Cocoa aroma
Astringent a b c
Sweetness Cocoa taste a a a
Bajo calima Sabaletas Cisneros
Basic flavors atributes Fine flavors atributes

Figure 11. Sensory attributes of chocolates produced with cacao beans from zone 1 (Cisneros), zone 2 (Bajo Calima) and zone 3 (Sabaletas). Different
letters indicate significant differences (ANOVA, Tukey's test, P < 0.050).
compounds such as phenolic compounds and methylxanthines that contribute to the formation of flavor precursor metabolites.
(theobromine, caffeine and theophylline) contribute to defining To monitor these processes, important physicochemical variables
the flavor profile and overall quality of cacao. such as internal cacao mass temperature and internal pH of the
Fermentation variables play a crucial role in the cacao seed's cacao beans were measured throughout the fermentation pro-
mass and heat transfer processes, which are closely tied to the cess (Fig. 9).
biochemical changes occurring within the seed. These changes Overall, an increase in temperature was observed in all regions,
12
involve a combination of enzymatic and non-enzymatic reactions albeit with slight variations in the trend. Zone 1 exhibited a

significant upward trend starting at 24 h, whereas, in zones 2 and distinct cacao notes. LAB has been shown to transform citric acid
3, the temperature increase began after 48 and 72 h, respectively. during the initial stages of fermentation, preventing its transfer
Beyond the 72-h mark, a consistent progressive temperature rise into the beans, and also reduces fructose to mannitol, which
was observed in all regions, eventually reaching approximately may diffuse into the beans.56-57 Recent evidence has demon-
45°C. This temperature increase is attributed to the exothermic strated the accumulation of mannitol during the fermentation
reactions caused by acetic acid bacteria thriving in the aerobic process. These transformations can lead to additional processes
conditions generated by the aeration of the fermentation mass. that contribute to the cacao flavor, beyond those caused by acetic
A concurrent sharp decrease in pH was observed throughout the acid alone.58
fermentation process across all zones. The pH levels ranged from
6.5 to 4.5. These pH conditions in all regions facilitate the optimal Sensory profile of chocolate and special quality
functioning of enzymatic processes and biochemical changes Chocolates made from the cacao beans of the three sampling
within the cacao seeds. Enzymes such as proteases, invertases zones exhibited distinct sensory attributes, including fresh fruits,
and polyphenol oxidases effectively act on proteins, carbohydrates floral notes, spices and nuts (Fig. 11). These unique attributes clas-
and phenolic compounds at the ideal temperature and pH, leading sify the cacao from this region as a fine flavor product, according
to the generation of lower molecular weight compounds known as to the International Cocoa Organization.59 Producers who are able
flavor precursor metabolites in cacao. These metabolites include to obtain higher quality cacao beans have greater opportunities
free amino acids, peptides and reducing sugars, all of which con- to access different market segments and achieve better profitabil-
tribute to the development of unique flavors in cacao. ity in their sales.39,54
Although there were differences in climate and genetic compo-
Microorganisms sition of the plantations, as well as variations within each fermen-
Bacterial diversity tation batch per zone, we did not find significant differences in
Cacao bean fermentation is a vital process that plays a key role in cacao bean quality among the zones. Some minor variations
the development of chocolate (or cocoa) flavor by generating in the intensity of certain attributes, such as fresh fruits, overall
essential chemical precursors. During fermentation, both bacteria quality and astringency, were observed (Fig. 11); however, there
and fungi undergo physical and biochemical transformations of were no distinct patterns in the types of attributes that formed
the bean structures. This transformation follows a predictable the sensory profiles of each region.
microbial succession, involving yeast and various functional Thus, there is no direct correlation between climate and the
groups of bacteria. Specifically, the bacterial succession starts with cacao varieties cultivated in these regions that determine the pro-
organisms related to the Enterobacteriaceae (ENT group), which duction of fine flavor cacao beans. Buenaventura demonstrates a
then transition to lactic acid bacteria (LAB group) and finally to combination of natural characteristics that contribute to the pro-
acetic acid bacteria (AAB group).54 The dynamics of microbial duction of high-quality cacao products, with a unique and spe-
communities involved in cacao bean fermentation can now be cialty quality associated with fruity-aroma chocolates.
assessed using high-throughput sequencing of phylogenetic
markers, a method known as metataxonomy. This approach over-
comes the limitations of traditional cultivation-based methods CONCLUSIONS
and provides a more detailed understanding of the microbial Colombia boasts a rich diversity of cultivated cacao throughout
composition involved in the fermentation process.55 the country, which is preserved in ex situ collections such as the
The fermentation processes in the three study zones were national germplasm collection conserved at AGROSAVIA research
examined using 16S rRNA metataxonomy (Fig. 10). Although all stations.31 This genetic reservoir plays a crucial role in the devel-
three microbial groups were observed in all fermentations, there opment of high-quality cacao in the country but, to enhance this
were differences in the fine-level composition, such as genus diversity, methodologies for DOs need to be implemented.
and species, as well as relative abundance, between the fermenta- In the present study, we propose an evidence-based framework
tions. For example, Tatumella was present during the initial fer- for supporting establishing DOs of cacao. Our analysis indicates spe-
mentation steps in zones 1 and 3 but not in zone 2, whereas cific GIs for each of the three sampling zones, with Buenaventura
Pantoea was only found in zone 3. Another unexpected difference identified as a region with natural characteristics to produce fine fla-
was the presence of Pseudomonas in zone 1, a bacterial taxo- vour cocoa products. Each zone exhibited a high differentiation and
nomic group that is not typically found in fermentations. Perhaps diversity of cacao cultivars. Buenaventura has the potential to be des-
the most notable difference between the fermentations was the ignated as a future DO for cacao from the Pacific region of Colombia,
atypical microbial succession observed in the fermentation from characterized by its unique fruity-aroma chocolates. However, each
zone 1, where it transitioned directly from ENT to AAB, bypassing study zone exhibited specific metabolomic characteristics influ-
the LAB group. This faster transition to AAB might be responsible enced by the unique climatic conditions of each zone. Additionally,
for the rapid increase in fermentation temperature and drop in pH distinct pest species and genetic footprints were identified across
observed at 48 h (Fig. 10) because the production of acetic acid by the zones. Overall, this proposed framework highlights the potential
AAB is an exergonic reaction. This accelerated transition may for supporting GIs and DOs in cacao. The integration of scientific
impact the transformation of the pulp and beans during fermen- expertise and community collaboration was vital in achieving a com-
tation and could explain the differences in the metabolomics pro- prehensive understanding of the cacao quality and characteristics
file observed between zone 1 and the other two locations. specific to Buenaventura.
Although recent studies have not found a clear relationship
between the absence of lactic acid bacteria (LAB) in cacao fermen-
tation and distinct characteristics in the organoleptic profile of AUTHOR CONTRIBUTIONS
chocolate, it is still possible that LAB play a role in the fine transfor- CEG-O and SE conceived the idea. CEG-O, SE, MP, RY, CR, MS, MZ,
13
mation of pulp and contribute to the development of sweet and AC-Q, MS and JR collected, cleaned and analyzed the data. CEG-O

and SE led the writing and revision of the manuscript. All authors indication for products commercialized in the international market.
contributed critically to the drafts and approved the final version Revista Ingi 4:1033–1047 (2020).
11 González-Orozco CE and Pesca A, Regionalization of Cacao (Theo-
of the manuscript submitted for publication. broma cacao L.) in Colombia. Front Sustainable Food Syst 6:1–11
(2022).
12 Pineda-Suarez P, Hacia una mejor coordinación interinstitucional para
ACKNOWLEDGMENTS una operación eficiente y completa de las entidades delegadas de
We thank the Corporación Colombiana de Investigación Agrope- las DOP en Colombia. Perspectivas mundiales de las indicaciones
cuaria (AGROSAVIA) for providing funding. We gratefully acknowl- geográficas Montpellier, Francia, 5–8 Julio (2022).
13 Superintendencia de Industria y Comercio. SIC, https://bit.ly/3BIQuZW
edge the generous assistance provided by cacao farmers Colombia Superintendencia de Industria y Comercio. Denominaciones
associations of Buenaventura: Asociación de Productores Campe- de Origen y Marcas no tradicionales. Documento Tecnico:1–76 https://
sinos de Cisneros, Asociación de Agricultores y Cacaoteros de Bajo bit.ly/3dMW7OK. Colombia. https://www.sic.gov.co/sites/default/files/
Calima, and Asociación de Productores y Transformadores de Denominaciones_Origen_Marcas_Tradicionales%20%20.pdf
14 Oberthür T, Läderach P, Posada H, Fisher MJ, Samper LF, Illera J et al.,
Zacarías-Sabaletas. We also thank the University of the Andes,
Regional relationships between inherent coffee quality and growing
especially Professor Bart Van Hoof, Maite Rosales, Juanita Duque environment for denomination of origin labels in Nariño and Cauca,
and Jesús Riascos, as well as the researchers from AGROSAVIA, Colombia. Food Policy 36:783–794 (2011).
Darwin Martínez, Eliseo Polanco and Rafael Novoa. Bancoldex 15 Decazy F, Avelino J, Guyot B, Perriot JJ, Pineda C and Cilas C, Quality of
and the Buenaventura Chamber of Commerce are acknowledged different Honduran coffees in relation to several environments.
J Food Sci 68:2356–2361 (2003).
for contributing resources for the development of this research 16 Unidad de Planificacion Rural Agropecuaria (UPRA) https://www.upra.
project. gov.co/uso-y-adecuacion-de-tierras/evaluacion-de-tierras/
zonificacion Colombia, 2021. www.dof.gob.mx/nota_detalle.php?
codigo=5449991&fecha=29/08/2016#gsc.tab=0. Mexico
CONFLICTS OF INTEREST 17 Hernandez CE and Granados L, Quality differentiation of cacao beans:
implications for geographical indications. J Sci Food Agric 101:3993–
The authors declare that they have no conflicts of interest.
4002 (2021).
18 Escobar S, Santander M, Vaillant F, Zuluaga M and Rodríguez J, Fine
cacao beans production: tracking aroma precursors through a com-
DATA AVAILABILITY STATEMENT prehensive analysis of flavour attributes formation. Food Chem 365:
The data that support the findings of this study are available from 1–11 (2021).
the corresponding author upon reasonable request. 19 Santander M, Vaillant F, Sinuco D, Rodríguez J and Escobar S, Enhance-
ment of fine flavour cacao attributes under a controlled postharvest
process. Food Res Int 143:110236 (2021).
20 González-Orozco CE, Osorio-Guarín JA and Yockteng R, Phylogenetic
SUPPORTING INFORMATION diversity of cacao (Theobroma cacao L.) genotypes in Colombia.
Supporting information may be found in the online version of this Plant Genet Resour: Characterization Utiliz 20:203–214 (2023).
21 Pérez-Escobar OA, Lucas E, Jaramillo C, Monro A, Morris SK, Bogarín D
article. et al., The origin and diversification of the hyperdiverse flora in the
Chocó biogeographic region. Front Plant Sci 10:1328 (2019).
22 Mesa-Sánchez ÓJ and Rojo-Hernández JD, On the general circulation
REFERENCES of the atmosphere around Colombia. Rev Acad Colomb Cienc Ex Fis
1 Marie-Vivien D and Casabianca F, Geographical indications: protection Nat 44:857–875 (2020).
of a name or a logo? A risky shift, in Actes de la conférence Perspec- 23 Fick SE and Hijmans RJ, Worldclim 2: new 1-km spatial resolution cli-
tives mondiales sur les Indicateurs Géographiques. FAO; CIRAD, mate surfaces for global land areas. Int J Climatol 37:4302–4315
Montpellier, pp. 31–33. Perspectives mondiales sur les indications (2017).
géographiques, 2022-07-05/2022-07-08, Montpellier (France) 24 Lembrechts J, Van den Hoogen J, Aalto J, Ashcroft M, De Frenne P,
(2022). Kemppinen J et al., Global maps of soil temperature. Global Chang
2 The World Intellectual Property Organization (WIPO) 2022. https:// Biol 28:3110–3144 (2021).
www.wipo.int/geo_indications/en/. Last accessed 27_09_2022 25 González-Orozco CE, Thornhill AH, Knerr NJ, Laffan SW and Miller JT,
3 The World Intellectual Property Organization (WIPO) Geographical Biogeographical regions and phytogeography of the eucalypts.
Indicators: an introduction. https://www.wipo.int/edocs/pubdocs/ Diversity Distrib 20:46–58 (2014).
en/wipo_pub_952_2021.pdf. (2021). Last accessed 27_09_2022. 26 Mahob RJ, Feudjio Thiomela R, Dibog L, Babin R, Fotso Toguem YG,
4 Institute for the National Denomination of Origin (INAO), Guide du Mahot Mahot H et al., Field evaluation of the impact of Sahlbergella
demandeur d'une appellation d'origine protégée (AOP) ou d'une singularis Haglund infestations on the productivity of different Theo-
indication géographique protégée (IGP) à l'exception des vins, bois- broma cacao L. genotypes in the southern Cameroon. J Plant Dis Prot
sons alcoolisées et boissons spiritueuses https://www.inao.gouv.fr/ 126:203–210 (2019).
eng/Official-signs-identifying-quality-and-origin/PDO-AOC (2017). 27 Reitan T, Nielsen, a, do not divide count data with count data; a story
5 Fouet O, Loor Solorzano RG, Rhoné B et al., Collection of native from pollination ecology with implications beyond. PloS One 11:
L. accessions from the Ecuadorian Amazon highlights a hotspot of e0149129 (2016).
cocoa diversity. Plants People Planet 4:618–637 (2022). 28 Cosme S, Cuevas HE, Zhang D, Oleksyk TK and Irish BM, Genetic diver-
6 González Jiménez E, Denominación de origen Cacao Chuao (fao.org). sity of naturalized cacao (Theobroma cacao L.) in Puerto Rico. Tree
FAO-IICA consultancy (2007). Genet Genomes 12:88 (2016).
7 Quingaísa, E, Riveros H, Estudio de Caso: Denominacion de Origen 29 Fang W, Meinhardt LW, Mischke S, Bellato CM, Motilal L and Zhang D,
“Cacao Arriba” (fao.org). (2007). Accurate determination of genetic identity for a single cacao bean,
8 Quintana Lombeida MD and Aguilar Herrera JV, Denominación de ori- using molecular markers with a nanofluidic system, ensures cacao
gen de cacao ecuatoriano: ¿Un aporte de marketing global? Innova authentication. J Agric Food Chem 62:481–487 (2014).
3:68–76 (2018). 30 Ji K, Zhang D, Motilal LA, Boccara M, Lachenaud P and Meinhardt LW,
9 SEGOB, Declaración General de Protección de la denominación de ori- Genetic diversity and parentage in farmer varieties of cacao (Theo-
gen Cacao Grijalva (2016), Mario Pomares: La denominación de broma cacao L.) from Honduras and Nicaragua as revealed by single
Origen garantiza la calidad de la Granada Mollar de Elche y la nucleotide polymorphism (SNP) markers. Genet Resour Crop Evol 60:
defiende frente a terceros. 441–453 (2013).
10 de Melo A, de Tarso P, de Souza S, de Melo C, da Conceição S and 31 Osorio-Guarín JA, Berdugo-Cely J, Coronado RA, Zapata YP,
Ribeiro A, Tomé-açu cacao: the importance of geographical Quintero C, Gallego-Sánchez G et al., Colombia a source of
14

cacao genetic diversity as revealed by the population structure 45 Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA,
analysis of germplasm bank of Theobroma cacao L. Front Plant Turnbaugh PJ et al., Global patterns of 16S rRNA diversity at a depth
Sci 8:1994 (2017). of millions of sequences per sample. Proc Natl Acad Sci U S A 108:
32 Wang B, Motilal LA, Meinhardt LW, Yin J and Zhang D, Molecular char- 4516–4522 (2011).
acterization of a cacao germplasm collection maintained in Yunnan, 46 Pacheco-Montealegre ME, Dávila-Mora LL, Botero-Rute LM, Reyes A
China using single nucleotide polymorphism (SNP) markers. Trop and Caro-Quintero A, Fine resolution analysis of microbial communi-
Plant Biol 13:359–370 (2020). ties provides insights into the variability of cacao bean fermentation.
33 Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, Front Microbiol 11:650 (2020).
McWilliam H et al., Clustal W and Clustal X version 2.0. Bioinformatics 47 Martin M, Cutadapt removes adapter sequences from high‐through-
23:2947–2948 (2007). put sequencing reads. EMBnet J, [S.l.] 17:10–12, May 2011. ISSN
34 Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S et al., 2226‐6089. Available: https://journal.embnet.org/index.php/
Geneious basic: an integrated and extendable desktop software embnetjournal/article/view/200 [11 October 2023]. https://doi.org/
platform for the organization and analysis of sequence data. Bioin- 10.14806/ej.17.1.200.
formatics 28:1647–1649 (2012). 48 Bolger AM, Lohse M and Usadel B, Trimmomatic: A flexible trimmer for
35 Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W and illumina sequence data. Bioinformatics (2014) btu170.
Gascuel O, New algorithms and methods to estimate maximum- 49 Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P et al., The
likelihood phylogenies: assessing the performance of PhyML 3.0. SILVA ribosomal RNA gene database project: improved data proces-
Syst Biol 59:307–321 (2010). sing and web‐based tools. Opens external link in new window. Nucl
36 Hammer Ø, Harper DAT and Ryan PD, PAST: paleontological statistics Acids Res 41:D590–D596 (2013).
software package for education and data analysis. Palaeontol Elec- 50 Mújica J and Martínez D, Manejo de plagas y enfermedades que afec-
tron 4:1–9 (2001). tan al cultivo del cacao. Instituto Colombiano Agropecuario (ICA)
37 Motamayor JC and Lachenaud P, da Silva E Mota JW, Loor R, Kuhn DN, (2015).
Brown JS, Schnell RJ, geographic and genetic population differenti- 51 Babin R, in Pest Management in Organic Cacao: Handbook of Pest Man-
ation of the Amazonian chocolate tree (Theobroma cacao L). PloS agement in Organic Farming, ed. by Vincenzo SK. CAB International,
One 3:e3311 (2008). Nairobi, Kenya pp. 502–518 (2018).
38 Alexander DH, Novembre J and Lange K, Fast model-based estimation 52 Bailey BA, Evans HC, Phillips-Mora W, Ali SS and Meinhardt LW, Moni-
of ancestry in unrelated individuals. Genome Res 19:1655–1664 liophthora roreri, causal agent of cacao frosty pod rot. Mol Plant
(2009). Pathol 19:1580–1594 (2018).
39 Escobar S, Santander M, Useche P, Contreras C and Rodríguez J, Align- 53 Jaimes YY, Ribeyre F, Gonzalez C, Rojas J, Furtado EL and Cilas C, Fac-
ing strategic objectives with research and development activities in tors affecting the dynamics of frosty pod rot in the main cocoa
a soft commodity sector: a technological plan for colombian cacao areas of Santander state, Colombia. Plant Dis 103:1665–1673
producers. Agri 10:2–32 (2020). (2019).
40 International Organization for Standardization, ISO, 13299:2016. Gen- 54 Lefeber T, Janssens M, Camu N and De Vuyst L, Kinetic analysis of
eral guidance for establishing a sensory profile. Geneva, strains of lactic acid bacteria and acetic acid bacteria in cacao pulp
Switzerland, Sensory analysis Methodology (2016). simulation media toward development of a starter culture for
41 Mayorga-Gross AL, Quirós-Guerrero LM, Fourny G and Vaillant F, An cacao bean fermentation. Appl Environ Microbiol 76:7708–7716
untargeted metabolomic assessment of cacao beans during fer- (2010).
mentation. Food Res Int 89:901–909 (2016). 55 Adler P, Bolten CJ, Dohnt K, Hansen CE and Wittmann C, Core fluxome
42 Thonusin C, IglayReger HB, Rothberg AE, Burant CF, Charles R and and metafluxome of lactic acid bacteria under simulated cacao pulp
Evans J, MetaboDrift, a drift‐correction and normalization tool for fermentation conditions. Appl Environ Microbiol 79:5670–5681
multi‐batch LC‐MS metabolomics data: application to study of (2013).
mixed‐meal tolerance in obese and non‐obese subjects. Chromatogr 56 De Vuyst L and Weckx S, The cacao bean fermentation process: from
A 1523:265–274 (2017). ecosystem analysis to starter culture development. J Appl Microbiol
43 Camu N, De Winter T, Verbrugghe K, Cleenwerck I, Vandamme P, 121:5–17 (2016).
Takrama JS et al., Dynamics and biodiversity of populations of lactic 57 De Vuyst L and Leroy F, Functional role of yeasts, lactic acid bacteria
acid bacteria and acetic acid bacteria involved in spontaneous heap and acetic acid bacteria in cacao fermentation processes. FEMS
fermentation of cacao beans in Ghana. Appl Environ Microbiol 73: Microbiol Rev 44:432–453 (2020).
1809–1824 (2007). 58 ICCO. (2017). Fine or Flavour Cocoa. Retrieved January 14, 2018, from
44 Faith JJ, Guruge JL, Charbonneau M, Subramanian S, Seedorf H, http://www.icco.org/about-cacao/fine-or-flavour-cacao.html
Goodman AL et al., The long-term stability of the human gut micro- 59 Beg MS, Ahmad S, Jan K and Bashir K, Supply chain and processing of
biota. Science 341:1237439 (2013). cacao—a review. Trends Food Sci Technol 66:108–116 (2017).
15


Gonzalez Orozcoetal2023

Uploaded by

Copyright:

Available Formats

Gonzalez Orozcoetal2023

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gonzalez Orozcoetal2023

Uploaded by

Copyright:

Available Formats

Research Article

(wileyonlinelibrary.com) DOI 10.1002/jsfa.13016

Integrating new variables into a framework to

Keywords: Buenaventura; chocolate; cocoa ﬂavour-aroma; rainforest

f Muséum National d'Histoire Naturelle, UMR-CNRS 7205, Paris, France

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

MATERIALS AND METHODS Geography

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

Table 1. Average climate conditions for the three sampling zones

Maximum Mean Minimum Vapor Wind

1. Cisneros 8.74 16.310 29.06 24.70 20.32 2686 2.52 1.05

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

Figure 2. Flow diagram of the proposed framework.

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

CHCM ETR O_7

C CN _5 tas _Po rve

SC C_1 10_1 La_P

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

The model demonstrated high predictive power, with a cross-

tions, leading to the generation of volatile aroma compounds

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

(A) (B) (C)

Basic flavors atributes Fine flavors atributes

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

wileyonlinelibrary.com/jsfa © 2023 The Authors. J Sci Food Agric 2023

J Sci Food Agric 2023 © 2023 The Authors. wileyonlinelibrary.com/jsfa

You might also like