Food Research International 132 (2020) 109061

Contents lists available at ScienceDirect

Food Research International

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

Nutritional value and antioxidant compounds during the ripening and after T
domestic cooking of bananas and plantains
⁎
C.V. Borgesa, , M. Maraschinb, D.S. Coelhob, M. Leonelc, H.A.G. Gomezd, M.A.F. Belina,
M.S. Diamantea, E.P. Amorime, T. Gianetia, G.R. Castroa, G.P.P. Limaa
a
São Paulo State University, Department of Chemistry and Biochemistry, Institute of Bioscience, 18.618-000 Botucatu, São Paulo, Brazil
b
Federal University of Santa Catarina, Plant Morphogenesis and Biochemistry Laboratory, 88.040-900 Florianopolis, Santa Catarina, Brazil
c
Center of Tropical Roots and Starches, CERAT, São Paulo State University, UNESP, 18.610-370 Botucatu, São Paulo, Brazil
d
Universidad Nacionalde Agricultura, Department of Food Technology, Barrio El Espino, Catacamas, Honduras
e
Embrapa Cassava & Fruits, 44.380-000 Cruz das Almas, Bahia, Brazil

A R T I C LE I N FO A B S T R A C T

Keywords: Genotypes of bananas and plantains have been studied for biofortification purposes, mainly due to content of
Phenolic compounds resistant starch (RS) and polyphenols. This study aims to identify banana and plantain genotypes with a high
Minerals content of resistant starch, phenolic compounds and minerals, and to evaluate the impact of the ripening stage
Resistant starch and domestic thermal processing to select superior genotypes with high levels of functional compounds. In this
Plant breeding
study, it was used bunches of bananas and plantain genotypes. The phenolic compounds profiles were de-
Thermal process
Musa spp.
termined by HPLC-DAD in pulps and peels. The resistant starch and the minerals (K, Na, Zn, Cu and Fe) were
evaluated in pulps and peels of unripe fruit. The results of phenolic compounds were studied in three ripening
stages, and after thermal processing (ripe stage) of two genotypes, which were most promising for biofortifi-
cation studies. Resistant starch and minerals were analysed in the unripe fruits. The peel biomass showed the
highest values of phenolic compounds and minerals. The total starch content in the pulp varied from 42.3%
(‘FC06-02′) to 80.6% (‘Pelipita’). Plantains and cooking bananas presented the highest contents of starch and
resistant starch (stage 2 – green with yellow traces). The pulps of the dessert genotypes ‘Khai’ and ‘Ouro da
Mata’, and cooking genotype ‘Pacha Nadam’ stood out due to their minerals high contents (P, K and Fe; Zn and
Fe; Ca, Mg and Zn, respectively). The dessert bananas (e.g., ‘Ney Poovan’) and cooking bananas (e.g., ‘Tiparot’)
had the highest concentrations of phenolic compounds, mainly in ripe fruit (stage 5 – yellow with green). In
addition, the thermal processing of Musa spp. fruit led to increasing these secondary metabolites, mainly the
cooking of fruit with peel by boiling, which should be preferred in domestic preparations.

1. Introduction
Today’s consumers tend to prefer ingesting foods that besides
nourishment, provide health benefits. Bananas and plantains are the
fourth most produced food items worldwide, with an annual production
of 107 million tons, mostly from India, China, the Philippines, Ecuador
and Brazil (FAO, 2017). These fruits stand out from other tropical fruit,
mainly due to their high consumption and versatility of use (raw, fried
and cooked, among others), besides being considered low-cost food and
having high energetic value (rich in starch).
Research on the genetic breeding of banana tree has identified and
explored genotypes with substantial amounts of resistant starch (RS)
and phytochemicals (Borges et al., 2014; Ghag & Ganapathi, 2018). The
banana and plantain fruits (Musa spp.) are considered to be a good
source of antioxidants, such as phenolic compounds, and minerals
(potassium and phosphorus), mainly in the peel (Pérez-Jiménez &
Saura-Calixto, 2018; Vu, Scarlett, & Vuong, 2019). Peel is the main
byproduct that has been described as a source of nutrients and anti-
oxidants compounds, and can be used by food producers in the forms of
flour and biomass (Khoozani, Birch, & Bekhit, 2019). In addition, mi-
cronutrients, such as zinc (Zn) and iron (Fe) are being widely re-
searched in biofortification programs (e.g., HarvestPlus), due to the
malnutrition problems associated with the lack of these minerals in less-
developed countries (Barrameda-Medina et al., 2017; Genc, Humphries,
Lyons, & Graham, 2005).
Many studies focus on the antioxidant, anti-ulcerogenic, anti-

⁎
Corresponding author.
E-mail address: cristine.vanzb@gmail.com (C.V. Borges).

https://doi.org/10.1016/j.foodres.2020.109061
Received 30 March 2019; Received in revised form 2 February 2020; Accepted 2 February 2020
Available online 03 February 2020
0963-9969/ © 2020 Elsevier Ltd. All rights reserved.
C.V. Borges, et al. Food Research International 132 (2020) 109061

hypertensive and anti-carcinogenic constituents in Musa spp. fruit, such
as vitamins A, B, C and E, β-carotene and phenolic compounds (ca-
techins, lignins, tannins and anthocyanins) (Borges et al., 2014, 2019;
Borges et al., 2018; Ghag & Ganapathi, 2018). In addition, the green
banana has been claimed as an important source of resistant starch
(RS), a non-digestible polysaccharide. RS can decrease the glycaemic
and insulinemic response, with important action in controlling the
metabolic syndrome (Sardá et al., 2016).
Research on the local cultivars is attractive because it takes into
account the vast genetic diversity of Musa spp. germplasms in the tro-
pical region. This diversity could be explored to identify suitable gen-
otypes for genetic breeding programs focused on the biofortification or
to expand the options for banana consumption. It is worth emphasising
that the content of phytochemicals in fruits, mainly the phenolic
compounds, vary significantly during their ripening stages and due to
the different types of domestic processing. Such inconsistency high-
lights the relevance of analysing these antioxidant compounds
throughout the ripening stages and after thermal processing of the fruit.
This study aimed to 1) characterize the banana and plantain geno-
types regarding their contents of RS, minerals and phenolic compounds
in pulps and peels, 2) evaluate the effect of ripening on the phenolic
compounds content, and 3) measure the retention of the phenolic
compounds after exposing the fruits (cooking bananas and plantains) to
thermal processing
2. Material and methods
2.1. Plant material
Bunches of bananas (22 genotypes) were harvested between March
and June of 2016, from different genomic groups belonging to the
Active Germplasm Bank (AGB) of banana maintained by Embrapa
Cassava & Fruits (Cruz das Almas, Bahia, Brazil). All the studied gen-
otypes were cultivated in the same place (latitude 12°40′12 ' S; long-
itude 39°06′07′' W; altitude 225 m) (Table 1). These genotypes (20
plants per genotype) were planted in three replications (n = 3). Field
planting was done in April 2015. When the branches reached ripening
stage 1, the second and third hands (each containing about 10 fingers)
of each genotype ( ± 2 bunches = 40 fruit) were harvested. All fruits
were stored at room temperature (20 ± 2 °C) and 80 ± 2% relative
humidity, until reaching ripening stage 7 (Borges et al., 2018).
2.2. Ripening stage determination
The fruit were divided into seven stages, according to their ripening
stages, considering the peel color, i.e., 1 = green; 2 = green with
yellow traces; 3 = more green than yellow; 4 = more yellow than
green; 5 = yellow with green; 6 = completely yellow; 7 = yellow with
coffee color areas (Soltani, Alimardani, & Omid, 2011). Fruit samples in
stage 2 of ripening were analysed for nutritional composition (starch,
RS, total phenolics [TP], antioxidant activity and minerals) and, to
study the effect of ripening stage on the TP content, the ripening stages
2, 5 and 7 (see Fig. 1, Supplementary material), were selected, which
are the most used for processing and/or raw consumption (Borges et al.,
2018). Further, the genotypes that stood out regarding their TP content
were also analysed by high-performance liquid chromato-
graphy–photodiode array detection (HPLC–DAD). The samples (pulp
and peels) were sliced, lyophilized and stored in an ultra-low tem-
perature freezer (-80 °C) (dry weight).
2.3. Study of the thermal processing impact on the phenolic compounds
In order to study the impact of the thermal processing on the TP
content, ripe fruit (stage 5) samples of the ‘D'Angola’ plantain, which is
already used commercially, and of the cooking banana ‘Pelipita’ were
used. The analyses were performed on peeled and unpeeled fruit. The
preparation methods included boiling fruits in water, microwaving and
stir-frying (with 2 mL soybean oil). Peeled and unpeeled bananas and
plantains (stage 5) were boiled in approximately 300 mL of water for
10 min in a covered stainless-steel pan. The remaining water was
drained off, and the fruit were cooled to room temperature. For the
microwave process, the whole banana and plantain were microwaved
(with and without peel) for 2 min, using a commercial microwave oven
(Panasonic, 21 l, output power of 700 W) (full power). The peel was
sliced longitudinally to avoid the increased internal steam pressure
inside of fruits. Stir-frying occurred for 5 min on samples of pulps sliced
longitudinally, using a stainless-steel frying pan greased with soybean
oil ( ± 2 mL). The thermal processing was performed with three re-
petitions (three fruits per repetitions). All analyses were performed in
triplicate. Furthermore, the pulps and peels were ground to a fine
powder (A.11, IKA, Germany) in liquid N2, lyophilised and stored at
–80 °C until the analysis (Borges et al., 2018). The TP content (spec-
trophotometry) and the profiles of the phenolic compounds (HPLC)
were analysed.

Table 1 2.4. Analyses of starch and RS in pulp of unripe fruits

Genotypes investigated of bananas and plantains.
Total starch (TS) content was determined according to Goñi, Garcia-
Genotypes Ploidy Subgroup/Subspecies Form of use
Alonso, and Saura-Calixto (1997) in unripe bananas and plantains (stg
Yangambi Km5 AAA Ibota Raw 2). The samples (50 mg) were suspended in 2 M KOH to disperse starch
Khai AAA Ibota Raw and then shaken at room temperature for 30 min, followed by in-
Pisang Kepok Bung AAB Peyan Raw
cubation (60 °C, 45 min, pH 4.75) with amyloglucosidase (1 mL of 300
Ney Poovan AB Ney Poovan Raw
Ouro da Mata AAAB * Raw U/mL; Sigma A-7255, USA) to hydrolyse the starch. The free glucose
Prata-Anã AAB Prata Raw was determined using glucose oxidase and peroxidase (Bergmeyer &
Grande Naine AAA Cavendish Raw Bernet, 1974). The TS was calculated as glucose × 0.9, and an analy-
Monthan 301 ABB Monthan Cooked tical standard of wheat starch (Sigma S-1514) was used as the reference
Monthan 172 ABB Monthan Cooked
for the analysis and calculations.
Simili Radjah ABB Peyan Cooked
Pelipita ABB Bluggoe Cooked The RS determination (Goñi, García-Diz, Mañas, & Saura-Calixto,
Pacha Nadan ABB Saba Cooked 1996) was performed in green bananas and plantains (stg 2) using
Namwa Khom ABB Pisang Awak Cooked 100 mg of samples passed through a sieve number 100-ABNT, followed
Muisa Tia ABB Pisang Awak Cooked
by the addition of 10 mL KCl–HCl buffer (0.2 M, pH 1.5) and 0.2 mL of
FC06-02 AABB Figo Cooked
Tiparot ABBB Klue Teparod Cooked pepsin solution (0.1 g of pepsin [Sigma P-7012] in 10 mL of KCl–HCl
D’Angola AAB Plantain Cooked buffer solution, pH 1.5). The samples were kept in a water bath (40 °C)
Curare Enano AAB Plantain Cooked with shaking for 60 min. After cooling to room temperature, 0.1 M
Terra Sem Nome AAB Plantain Cooked Tris–maleate buffer (9 mL, pH 6.9, 0.1 M) and α-amylase (1 mL, 4 g α-
Tipo Velhaca AAB Plantain Cooked
amylase [Sigma A-3176] in 100 mL of Tris–maleate buffer) were added.
Terra Anã Branca AAB Plantain Cooked
Samurá B AAB Plantain Cooked The sample was incubated in a water bath at 37 °C for 16 h, under
stirring, collected, filtered through 110 mm filter paper and the filtrate

2
C.V. Borges, et al.
Fig. 1. Mean and standard deviation of total phenolic (TP) compounds of pulp (stage 2, 5, and 7). Genotypes separated by dessert bananas (A), cooking bananas (B),
and plantain (C).
discarded. The residue collected in the filter paper was transferred to a
50 mL tube, and 3 mL each of distilled water and 2 M KOH were added.
After holding for 30 min, with occasional shaking, 4.5 mL of 1 M HCl,
3 mL of sodium acetate buffer (pH 4.75) and 80 μL of amyloglucosidase
(0.144 g Sigma A-7255 amyloglucosidase in 10 mL water) were added.
The sample was incubated in a water bath at 60 °C (45 min), under
stirring and further hydrolysates were recovered by filtration (11.0/
12.5 filter paper). Glucose oxidase (2 mL) was added to each sample
(20 µL). The tubes were capped and held in a water bath at 37 °C for
10 min. Once cooled, the absorbance was read at 505 nm. RS quanti-
fication (%) was calculated by multiplying the calculated value of free
glucose by 0.9.
2.5. Mineral analysis
Sample digestion applied to fruit samples (pulp and peel, stg 2) was
3
Food Research International 132 (2020) 109061
adapted from Altundag and Tuzen (2011) with some modifications.
Briefly, a small aliquot of each sample, 1.0 g, (after lyophilization step)
was transferred to tared vessels liners (100 mL capacity). At each re-
action flask 8 mL of digestion solution were added. Digestion solution is
composed of concentrated HNO3-H2O2 in the proportion of 6:2 v/v. The
mixture (digestion solution + sample) remained under contact for 12 h
without heating.). The digestion conditions were: 2 min for 250 W,
6 min for 250 W, 5 min for 400 W, 8 min for 550 W, vent.: 8 min. After
cooling, the acid sample extract was transferred to 10 mL volumetric
flask. Blanks were prepared at the same way without sample. Three
aliquots of each fruit sample were subjected to acid digestion and metal
determination was performed at each aliquot (n = 3). The determi-
nation of K, sodium (Na), Zn, copper (Cu) and Fe was performed by
atomic absorption spectrophotometry using a Shimadzu AA 7000
(Japan) apparatus configured to use the more intense resonance lines
for each mineral. K and Na were determined by atomic emission, and
C.V. Borges, et al.
the microelements Cu, Zn and Fe were determined by atomic absorp-
tion using their respective hollow cathode lamps. For the calibration
curves, standard mineral solutions (1000 mg/L; Specsol, Brazil) were
used after step-wise dilution to the desired concentration.
2.6. Extraction of phenolic compounds and antioxidant activity
The lyophilised and powdered samples (pulps and peels, 500 mg)
were added of 2.5 mL MeOH: phosphoric acid: water (80: 0.1: 19.9, v/
v/v) solution, vortexed (1 min), and sonicated (2.4 GHz, 25 °C, 30 min -
Q3.0/40A, Ultronique, Brazil). Following a centrifugation step
(3800×g, 10 min - Hettich Zentrifugen, Mikro220R, Germany), the
extracts were recovered by collecting the supernatant, which were
transferred to amber tubes, and the residual pellets were submitted to
an additional extraction, as described above. The supernatants were
pooled and concentrated under vacuum in an Eppendorf concentrator
plus centrifugeTM (20 mbar, 1400 rpm, 30 °C – Eppendorf AG,
Hamburg, Germany). For further analysis, the samples were re-
suspended according to described in 2.8.1.
2.7. Analysis by UV–Vis spectrophotometry
2.7.1. Total contents of phenolic compounds
The TP (pulp and peel) was measured using the Folin–Ciocalteau
reagent (Singleton & Rossi, 1965), in 0.1 mL of extract. The values were
calculated from a standard curve of gallic acid, and the values were
expressed as milligrams (mg) of gallic acid equivalents (GAE) per 100 g
of dry weight (d.w.).
2.7.2. DPPH (2,2-diphenyl-1-picrylhydrazyl) radical scavenging assay in
pulp
The capacity of samples (extracts) to reduce the DPPH radical was
assessed using the method of Brand-Williams, Cuvelier, and Berset
(1995). Pulp extracts (500 µL) were allowed to react with 300 µL of the
DPPH solution for 60 min in the dark condition. The absorbance was
read at 515 nm in a UV–Vis Ultrospec 3000 spectrophotometer (Phar-
macia Biotech, Uppsala, Sweden), and the results were expressed as mg
Trolox equivalents (TE) from the standard curve (y = 0.6992x −
0.0044, R2 = 0.99) per 100 g (d.w.).
2.7.3. FRAP (ferric-reducing antioxidant power) assay in pulp
The antioxidant activity was determined with the ferric-reducing
ability assay (Benzie & Strain, 1996). In brief, 900 µL of the fresh FRAP
reagent (300 nM acetate buffer, pH 3.6; 2.5 mL of 10 mM 2,4,6-tris(2-
pyridyl)-5-triazine [TPTZ] in 40 mM HCl; 2.5 mL FeCl3 solution) was
added to 30 µL of extract and incubated for 30 min in the dark con-
dition. The absorbance was measured at 595 nm. The absorbance va-
lues were compared to a calibration curve of ferrous sulphate (FeSO4)
and expressed as millimoles of Fe equivalents per 100 g (d.w.).
2.7.4. ABTS (2,2-azino-bis-(3-ethylbenzothiazoline-6-sulphonic acid)
radical scavenging activity assay in pulp
The antioxidant activity was also determined by the sequestration of
the ABTS radical, according to Re et al. (1999). The absorbance was
measured in a UV–Vis Ultrospec 3000 spectrophotometer (Pharmacia
Biotech) at 734 nm, and the results were expressed as millimoles of TE
per 100 g (d.w.).
2.8. Reversed-phase- HPLC analysis
2.8.1. Phenolic compounds
The pulp extract (section 2.6) was solubilised in 500 µL of mobile
phase [250 µL of solvent A (water: trifluoracetic acid, 99.9:0.1, v/v)
and 250 µL of solvent B (100% acetonitrile)] and centrifuged (5 min,
5000 × g). The supernatant was recovered, and an aliquot (10 µL) was
injected into the HPLC system (Ultimate 3000, Thermo Scientific,
4
Food Research International 132 (2020) 109061
Bremen, Germany) coupled to a DAD and C18 column (150 × 4.6 mm,
Kinetex® 2.6 µm F5 100 Å; Phenomenex, USA). The chromatographic
elution was made using the following gradient: 90% A and 10% B
(5 min), 60% A and 40% B (5 min), 30% A and 60% B (3 min), 90% A
and 10% B (2 min), at a flow rate of 0.75 mL min−1. The phenolic
compounds were measured at 270 nm (gallic acid), 254 nm (hydro-
xybenzoic acid), 270 nm (epigallocatechin gallate and catechin) and
254 nm (quercetin). by calculating the peak areas of interest.
Calibration curves were built using known amounts of analytical stan-
dards (99.98% purity; Sigma–Aldrich Co., St. Louis, MO, USA). The
results were expressed as micrograms per gram of sample (d.w.), taking
into account the average of three consecutive injections per sample
(n = 3).
2.9. Statistical analysis
The experiments were carried out following an entirely randomized
design. The data set of TS, RS, TP, DPPH, FRAP, ABTS, and mineral
contents were collected only from the unripe fruit samples, summarized
and submitted to one-way analysis of variance (one-way ANOVA),
followed by the Scott-Knott’s test (p < 0.05) for mean comparison
among the genotypes, in triplicate. On the other hand, the data of total
phenolic contents recorded from the ripening stage samples were ana-
lysed in a factorial scheme (22 × 3, genotypes × ripeness) and sub-
mitted toone-way ANOVA, followed by the Scott-Knott’s and Tukey’s
tests (p < 0.05) for comparing treatment means among the genotypes,
and the ripening stages, respectively. The data collected after thermal
processing treatments were also submitted to one-way ANOVA, fol-
lowed by the Tukey’s test (p < 0.05). All analyses were performed
with three repetitions (three fruit/experimental unit), and in triplicate.
The one-way ANOVA and PCA of the data sets were performed using
the algorithms of the statistical softwares SISVAR (Ferreira, 2011) and
XLSTAT (v. 2017, Addinsoft, France), respectively. PCA aimed at to
detect possible correlations among the variables from the chemical
profiles and the ripening stages, as well as from the TP and the geno-
types after thermal processing treatments.
3. Results and discussion
3.1. Starch, resistant starch, total phenolic compounds and minerals in
unripe bananas and plantains
Fruit samples in stage 2 of ripening were analysed for nutritional
composition. The TS contents varied from 42.3% (‘FC06-02′) to 80.6%
(‘Pelipita’), as the RS spanned over 22.9% (‘Namwa’) to 49.9% (‘Terra
Anã Branca’). A clear trend regarding the contents of those variables in
the genotypes investigated was not detected, since a wide range of
values was observed either for banana and plantain samples.
Interestingly, the genotypes ‘Terra Anã Branca’ (RS: 49.9%), ‘Pelipita’
(48.1%) e ‘Simili Radjah’ (48.2%) (Table 2) presented superior RS va-
lues to the genotype ‘D'Angola’ (41.2%), one of the most important
plantains in the Brazilian market. Similarly, this trait was noted in the
cooking bananas ‘Monthan 301′ (45.5%) and ‘Monthan 172′ (47.6%).
Among the dessert bananas, ‘Prata-Anã’ (47.5%) stood out for its high
RS content, followed by ‘Pisang K. Bung.’ (39.7%), especially when
compared to ‘Grande Naine’ (27.5%), the most marketed genotype
worldwide.
Studies have demonstrated that the RS content of green bananas
varies according to the variety and, when compared with Cavendish
bananas (e.g., ‘Grande Naine’ and ‘Nanicão’), which are used for the
production of banana flour in Brazil, plantain has a higher content of
total RS (Bi et al., 2019). It is noteworthy that the dessert and cooking
banana genotypes with superior RS quantities were detected in the
Musa spp. germplasm, mainly when compared with the commercially-
available cultivars. These results could be explored for increasing the
contents of those bioactive compounds through breeding programs of
C.V. Borges, et al. Food Research International 132 (2020) 109061

Table 2
Contents of total starch (TS), resistant starch (RS) and antioxidant activity in pulp samples (stg 2), total phenolic (TP) in peel samples (stg 2) and moisture content in
pulp of Musa spp. genotypes.
Genotypes TS (%) RS (%) TP peel (mg GAE/ DPPH pulp (mg TE/ FRAP pulp (mmol Fe/ ABTS pulp (mmol TE/ Moisture content
100 g d.w.) 100 g d.w.) 100 g d.w.) 100 g d.w.) (pulp) (%)

Dessert bananas
Yangambi K. 67.6 ± 4.3a 27.5 ± 2.4e 544.1 ± 7.3e 344.4 ± 3.4e 19.0 ± 7.0f 503.5 ± 16.2j 66.9 ± 0.8a
Khai 56.6 ± 5.1b 26.7 ± 2.3e 482.8 ± 6.5g 442.6 ± 3.9d 15.9 ± 7.4g 587.5 ± 18.9i 67.9 ± 0.4a
Pisang K. Bung 66.6 ± 5.3a 39.7 ± 0.4b 486.5 ± 7.4g 860.9 ± 9.6b 6.0 ± 3.2j 174.8 ± 18.3m 40.3 ± 1.1h
Ney Poovan 44.7 ± 1.6c 27.2 ± 0.2e 654.4 ± 7.4c 713.9 ± 7.1c 39.7 ± 5.1b 3650.9 ± 26.0a 60.1 ± 0.4c
Ouro da Mata 63.7 ± 4.2a 32.5 ± 1.9d 532.5 ± 7.6f 481.8 ± 4.9d 21.9 ± 0.9e 1008.4 ± 24.9f 63.8 ± 1.2b
Prata-Anã 72.0 ± 3.7a 47.5 ± 2.5a 374.2 ± 2.7k 326.5 ± 3.3e 15.7 ± 1.9g 1256.8 ± 4.4d 61.0 ± 0.9c
Grande Naine 54.7 ± 2.3b 27.5 ± 1.4e 409.1 ± 1.3j 256.8 ± 2.6e 18.5 ± 8.6f 529.7 ± 5.2j 66.4 ± 0.5a
Cooking bananas
Monthan 301 67.1 ± 4.6a 45.5 ± 0.4a 462.8 ± 7.5h 571.2 ± 5.7c 31.5 ± 5.2c 1480.8 ± 22.1c 67.8 ± 1.1a
Monthan 172 59.9 ± 2.8b 47.6 ± 0.4a 463.0 ± 7.3h 266.3 ± 2.7e 1.6 ± 2.5k 224.9 ± 26.1l 57.0 ± 1,7e
Simili Radjah 73.3 ± 4.4a 48.2 ± 0.6a 467.5 ± 7.5h 844.8 ± 9.4b 8.5 ± 3.0j 350.2 ± 20.6k 63.7 ± 0.4b
Pelipita 80.6 ± 1.2a 48.1 ± 2.2a 519.1 ± 7.5f 448.0 ± 4.5d 25.3 ± 5.9d 886.5 ± 33.0g 55.8 ± 0.4e
Pacha Nadam 66.4 ± 3.6a 38.2 ± 2.0c 534.0 ± 7.6f 442.5 ± 4.2d 20.4 ± 4.8e 800.1 ± 29.0h 62.6 ± 2.0b
Namwa 59.1 ± 1.4b 22.9 ± 0.8f 557.9 ± 1.5e 190.3 ± 1.9f 14.3 ± 4.3g 449.9 ± 14.2j 62.4 ± 0.6b
Muisa Tia 73.2 ± 1.5a 32.6 ± 2.3d 647.4 ± 2.6c 605.2 ± 6.0c 6.7 ± 1.4j 450.9 ± 23.6j 58.8 ± 2.1d
FC06-02 42.3 ± 2.3c 32.4 ± 2.9d 439.9 ± 6.5i 140.4 ± 1.4f 0.6 ± 0.1k 104.5 ± 20.5m 68.7 ± 0.7a
Tiparot 53.3 ± 1.3b 40.9 ± 2.6b 464.4 ± 5.8h 999.9 ± 0.3.8a 74.9 ± 2.2a 3494.6 ± 6.4b 58.9 ± 0.8d
Plantains
D’Angola 63.9 ± 3.1a 41.2 ± 1.4b 615.9 ± 3.2d 295.8 ± 2.9e 7.9 ± 2.1j 451.5 ± 27.5j 56.9 ± 0.5e
Curare Enano 63.0 ± 1.5a 39.7 ± 2.2b 655.4 ± 2.6c 972.5 ± 9.7b 13.9 ± 0.5g 1181.3 ± 25.2e 56.7 ± 0.8e
Terra S. N. 63.3 ± 2.8a 45.3 ± 0.5a 703.0 ± 4.1a 648.5 ± 6.5c 9.5 ± 0.7i 765.1 ± 22.2h 55.8 ± 1.2e
Tipo Velhaca 79.0 ± 3.8a 42.4 ± 1.3b 687.7 ± 2.8b 640.4 ± 6.1c 11.9 ± 2.1h 499.6 ± 17.4j 58.7 ± 1.4f
Terra A. B 68.2 ± 2.3a 49.9 ± 3.4a 676.8 ± 2.6b 532.3 ± 5.3d 7.3 ± 0.2j 250.5 ± 20.7l 58.7 ± 1.4d
Samurá B 63.3 ± 1.1a 45.6 ± 1.5a 615.0 ± 3.3d 718.1 ± 7.5c 20.9 ± 3.5e 651.4 ± 10.4i 51.6 ± 1.8g

*The same lower case letters (genotypes) and uppercase letters (ripening stages) do not differ by Scott and Knott test (5%) and Tukey test (5%), respectively. TS: total
starch; RS: resistant starch; TP: total phenolic.

the species or by introducing the genotypes into ongoing agricultural
systems.
The highest averages of TP compounds in the pulp were observed in
the dessert bananas and cooking bananas (Fig. 1). Superior values of
these antioxidant compounds in dessert bananas (247 mg/100 g fresh
weight) were also reported in other studies with Musa spp. (Tsamo
et al., 2014). This study showed that the TP content is one of the most
important attributes to differentiate dessert bananas (genomic con-
stitution A) from the cooking ones (non-plantains—genomic constitu-
tion B). However, in the present study, besides the high content of these
secondary metabolites found in the dessert bananas, the non-plantain
cooking bananas, such as ‘Pelipita’ and ‘Tiparot’, also presented such
trait, with no distinct division in the content of these compounds re-
garding subgroup and or mode of consumption.
Previous findings of our research group indicated that banana peels
contain various phenolic antioxidants in quantities higher than that
found in the banana pulps (Khoozani et al., 2019). The TP contents in
the peels varied from 374 mg (‘Prata-Anã’) to 703.0 mg GAE/100 g
(‘Terra S.N.’) (Table 2). The cooking bananas ‘Tiparot’ (509.74 mg/
100 g GAE) and ‘Pelipita’ (506.12 mg/100 g GAE) stood out among the
genotypes analysed (Table 2).
The mineral elements are involved in several vital functions of the
human body. The concentration of minerals in pulps and peels are
presented in Tables 3 and 4. It was noticed that in general, peels showed
higher contents than pulps, as described by Sulaiman et al. (2011). The
edible part of the banana fruit is considered a good source of K in the
human diet (Anyasi, Jideani, & Mchau, 2018). According to the dietary
reference intake (DRI) for adults, daily ingestion of 4700 mg of K is
recognised as adequate (IOM, 2000; Wall, 2006). In this context, 100 g
fresh weight (f.w.) of pulp of the analysed genotypes would provide
8.5% (‘Khai’) or 6.22% (‘Terra Sem Nome’) of the recommended DRI of
K ingestion, respectively. These contents are similar to those reported
for bananas from Malaysia that provided between 10% of the daily K
(Sulaiman et al., 2011).
Phosphorus (P) was the second most abundant mineral detected in
the pulps and peels, highlighting ‘Khai’ pulp (251.9 mg/100 g) and
‘Tiparot’ peel (350.4 mg/100 g) (Table 3 and 4). Taking into con-
sideration that the DRI for P is 700 mg, the consumption of 100 g (f.w.)
of ‘Khai’ pulp would provide 11.6% of the DRI for adults, which is
superior to that reported by Sulaiman et al. (2011) and Anyasi et al.
(2018). It is recognised that besides the genotype, other factors influ-
ence the minerals content, such as the abiotic factors and the agri-
cultural practices (Anyasi et al., 2018).
Magnesium (Mg) also presented relevant amounts (Table 3 and 4)
when compared with other minerals in the samples studied. ‘Pacha
Nadam’ pulp (166.56 mg/100 g) and ‘Tiparot’ peel (334.9 mg/100 g)
presented the highest contents (Table 3). These values are superior to
other Musa spp. genotypes studied (Sulaiman et al., 2011) and are si-
milar to the values found in genotypes originating from South African
communities (Anyasi et al., 2018). According to the results, 100 g (f.w.)
pulp of ‘Pacha Nadam’, for example, would provide 23% of the total
DRI of Mg for women (320 mg) and 18% of that for adult men (400 mg)
(Sulaiman et al., 2011). ‘Pisang K. B.’ pulp (120.6 mg/100 g) stood out
among the others, and ‘Prata-Anã’ peel presented values of 194.7 mg/
100 g, higher than the contents found in the other analysed genotypes.
Among the evaluated microelements, Zn and Fe have been widely
investigated in biofortification programs (e.g., HarvestPlus) due to the
malnutrition associated with the deficiencies in these minerals, mainly
in developing countries. The lack of a standard procedure to measure
the deficiency of Zn hinders the estimates of the number of people with
a Zn deficiency, but about 20% of the world population is thought to be
at risk (Barrameda-Medina et al., 2017). In the present study, higher Zn
contents were detected in the peel than pulp, mainly in the dessert
banana ‘Prata-Anã’ (6.20 mg/100 g) and ‘Pisang K. B.’ (6.10 mg/100 g),
which stood out from the others. In the pulps, a wide range of contents
was found, varying from 0.70 (‘Simili R.’) to 1.54 mg/100 g (‘Pacha
Nadam’). The cooking banana ‘Pacha Nadam’ (1.54 mg/100 g) and the
dessert banana ‘Ouro da Mata’ (1.53 mg/100 g) showed the highest
contents of this microelement (Table 3).
The microelement found in the highest content was Fe, present
mainly in the fruit peels, highlighting the dessert banana ‘Khai’ which
has the highest Fe content (Table 4). The pulp of the dessert banana

5
C.V. Borges, et al. Food Research International 132 (2020) 109061

Table 3
Mineral contents (mg/100 g d.w.) in pulp samples (stg 2) of Musa spp. genotypes.
Genotypes P K Na Ca Mg Fe Cu Zn

Dessert bananas
Yangambi K. 195.4 ± 3.9d 1006.7 ± 1.6b 43.7 ± 0.6g 52.9 ± 1.4d 118.0 ± 1.2d 18.1 ± 0.2b 0.34 ± 0.01a 1.03 ± 0.02e
Khai 251.9 ± 3.2a 1244.9 ± 2.4a 61.6 ± 1.6b 44.4 ± 1.7e 120.6 ± 1.8d 23.9 ± 0.7a 0.25 ± 0.01c 1.24 ± 0.03c
Pisang K. Bung 178.7 ± 4.5e 758.8 ± 3.3h 71.3 ± 1.1c 70.9 ± 1.4b 120.6 ± 1.9d 20.0 ± 1.4b 0.26 ± 0.02c 1.18 ± 0.03c
Ney Poovan 225.2 ± 2.3c 807.3 ± 2.6f 53.6 ± 0.7e 39.9 ± 2.0f 127.3 ± 1.4c 18.8 ± 1.4b 0.22 ± 0.03c 1.17 ± 0.02c
Ouro da Mata 190.8 ± 3.1d 996.3 ± 1.2c 53.2 ± 1.3e 47.1 ± 1.4e 133.5 ± 2.3b 23.5 ± 1.4a 0.20 ± 0.01d 1.53 ± 0.03a
Prata-Anã 167.3 ± 3.2f 764.9 ± 1.5g 45.8 ± 1.6g 38.9 ± 1.6f 129.2 ± 1.4c 19.8 ± 0.4b 0.27 ± 0.02c 1.01 ± 0.01e
Grande Naine 135.3 ± 1.7i 961.3 ± 2.8d 48.2 ± 2.9f 38.9 ± 1.4f 131.3 ± 1.3c 19.4 ± 1.8b 0.37 ± 0.01a 1.37 ± 0.02b
Cooking bananas
Monthan 301 236.0 ± 2.9b 829.1 ± 4.0e 43.9 ± 1.9g 43.5 ± 0.9e 121.7 ± 0.7d 18.9 ± 1.4b 0.29 ± 0.01b 1.36 ± 0.05b
Monthan 172 132.8 ± 0.1i 648.9 ± 4.3l 60.9 ± 2.7b 46.4 ± 1.7e 127.2 ± 1.3c 17.3 ± 0.4b 0.18 ± 0.02d 1.03 ± 0.03e
Simili Radjah 153.3 ± 4.0g 752.1 ± 4.2h 37.4 ± 0.5h 36.9 ± 1.4f 118.8 ± 1.3d 15.9 ± 0.3c 0.12 ± 0.01f 0.70 ± 0.05g
Pelipita 134.0 ± 2.8i 684.7 ± 2.2j 37.1 ± 1.9h 25.5 ± 1.8g 93.4 ± 1.1e 15.9 ± 1.4c 0.12 ± 0.01f 0.83 ± 0.01f
Pacha Nadam 234.6 ± 2.8b 647.8 ± 1.3l 50.7 ± 0.1f 76.0 ± 1.5a 166.6 ± 1.4a 23.7 ± 1.1a 0.17 ± 0.01d 1.54 ± 0.03a
Namwa 127.0 ± 3.2i 591.2 ± 2.8m 82.7 ± 3.3a 39.3 ± 3.3f 95.5 ± 1.4e 20.3 ± 1.2b 0.30 ± 0.02b 0.90 ± 0.05f
Muisa Tia 130.4 ± 2.6i 652.2 ± 4.1l 59.9 ± 2.1d 43.9 ± 1.4e 85.4 ± 1.6g 20.5 ± 2.8b 0.10 ± 0.01f 1.43 ± 0.01b
FC06-02 199.6 ± 2.3d 765.3 ± 4.1g 56.3 ± 1.4e 57.5 ± 1.0c 136.5 ± 2.8b 20.1 ± 1.7b 0.15 ± 0.02e 1.33 ± 0.03b
Tiparot 141.7 ± 3.4h 655.6 ± 2.4k 56.7 ± 2.1e 36.8 ± 1.3f 121.4 ± 1.6d 18.9 ± 1.7b 0.24 ± 0.03c 1.11 ± 0.04d
Plantains
D’Angola 119.7 ± 3.4j 702.5 ± 3.2i 93.3 ± 0.4a 22.2 ± 1.2g 96.9 ± 1.4e 19.2 ± 1.2b 0.14 ± 0.01e 1.02 ± 0.01e
Curare Enano 115.5 ± 3.7k 647.6 ± 2.9l 48.0 ± 2.8f 23.9 ± 1.4g 98.2 ± 1.4e 19.1 ± 1.4b 0.10 ± 0.02f 1.09 ± 0.10d
Terra S. N. 115.6 ± 2.4k 661.8 ± 1.2k 62.7 ± 2.1b 24.8 ± 1.4g 96.3 ± 1.4e 17.9 ± 1.4b 0.04 ± 0.01g 0.98 ± 0.03e
Tipo Velhaca 110.7 ± 2.5k 690.2 ± 4.9j 47.1 ± 1.4f 24.3 ± 2.9g 91.4 ± 1.7f 21.3 ± 1.2b 0.11 ± 0.01f 0.88 ± 0.04f
Terra A. B 112.1 ± 2.7k 686.5 ± 1.5j 55.8 ± 1.1e 23.4 ± 1.6g 95.5 ± 0.7e 19.8 ± 1.9b 0.25 ± 0.02c 0.75 ± 0.03g
Samurá B 121.1 ± 2.8j 646.2 ± 5.7l 45.2 ± 1.4g 22.3 ± 1.7g 98.3 ± 1.4e 19.3 ± 1.3b 0.22 ± 0.03c 0.99 ± 0.01e

*The same lower case letters do not differ by Scott and Knott test (5%).

‘Ouro da Mata’ (23.5 mg/100 g) and cooking banana ‘Pacha Nadam’
(23.7 mg/100 g) excelled regarding the Fe contents (Table 3). About 1/
3 of the world’s population has a Fe deficiency, and the World Health
Organisation (WHO) estimates that most pre-school aged children and
pregnant women in developing countries present a Fe deficiency (Genc
et al., 2005). Thus, both peel and pulp of these bananas genotypes can
be used as Fe supplementation in the diet.
3.2. Correlation analysis among the phenolic compounds, minerals and
antioxidant activity in Musa spp. fruit
In a second approach, multivariate statistical analysis was applied to
the data set of phenolic and mineral contents and antioxidant activity.
For that, a classification model was built, by adopting the principal
component analysis. The results showed that PC1 and PC2 explained
60.13% of the data variance (Fig. 2). The TP content showed the
highest correlation with antioxidant activity (DPPH: r = 0.4; FRAP:
r = 0.7; ABTS: r = 0.6, p < 0.05) grouped into in PC1 + and PC2+,
and not significant correlation with mineral content (P: r = 0.2; Mg: r

Table 4
Mineral contents (mg/100 g d.w.) in peel samples (stg 2) of Musa spp. genotypes.
Genotypes P K Na Ca Mg Fe Cu Zn

Dessert bananas
Yangambi K. 248.5 ± 3.2d 2224.0 ± 2.8n 62.1 ± 2.9f 180.5 ± 2.1h 152.4 ± 2.9j 26.9 ± 0.1e 0.9 ± 0.05b 2.9 ± 0.04g
Khai 220.4 ± 3.5e 3065.6 ± 2.3e 90.2 ± 2.5c 299.6 ± 3.5a 145.7 ± 2.9k 35.8 ± 0.2a 1.1 ± 0.06a 4.8 ± 0.02c
Pisang K. Bung 256.3 ± 2.6c 2670.6 ± 2.2k 157.7 ± 2.6a 190.5 ± 2.1g 290.7 ± 3.2b 30.0 ± 0.1c 0.5 ± 0.07d 6.1 ± 0.06a
Ney Poovan 247.8 ± 2.4d 2862.7 ± 3.5h 60.5 ± 4.6f 239.5 ± 2.1e 216.9 ± 2.8e 23.4 ± 0.6h 0.5 ± 0.03d 5.1 ± 0.08b
Ouro da Mata 204.6 ± 3.2f 2216.0 ± 2.7n 77.3 ± 2.5d 159.7 ± 3.5b 184.4 ± 3.8h 27.6 ± 0.1e 0.7 ± 0.05c 2.5 ± 0.03g
Prata-Anã 243.2 ± 3.3d 2839.9 ± 5.1i 71.2 ± 4.7e 277.4 ± 3.5b 194.7 ± 3.2g 25.9 ± 1.5f 0.3 ± 0.04e 6.2 ± 0.02a
Grande Naine 311.8 ± 2.6b 3243.6 ± 5.2c 104.5 ± 3.3b 253.5 ± 3.5d 172.8 ± 3.8i 33.8 ± 1.6b 0.7 ± 0.04c 4.7 ± 0.07d
Cooking bananas
Monthan 301 134.2 ± 2.2j 2077.8 ± 2.2o 57.7 ± 1.7f 211.6 ± 2.8f 186.3 ± 3.2h 24.9 ± 0.2g 0.8 ± 0.09c 3.8 ± 0.03f
Monthan 172 218.2 ± 3.2e 2046.1 ± 4.7p 55.8 ± 4.1f 200.8 ± 3.4g 190.3 ± 4.3g 26.2 ± 0.3f 0.7 ± 0.05c 2.8 ± 0.07g
Simili Radjah 213.4 ± 3.3e 2648.9 ± 4.6l 75.6 ± 3.1d 250.7 ± 2.8d 191.9 ± 3.4g 27.6 ± 0.1e 0.5 ± 0.03d 4.9 ± 0.04b
Pelipita 263.4 ± 4.0c 3075.1 ± 4.2e 64.9 ± 1.4e 107.8 ± 1.4k 200.6 ± 2.7f 25.9 ± 1.4f 0.7 ± 0.06c 2.2 ± 0.03i
Pacha Nadam 174.3 ± 2.6g 1575.0 ± 3.1s 60.4 ± 2.5f 261.4 ± 1.4c 232.8 ± 4.2d 24.0 ± 0.2h 0.7 ± 0.07c 2.9 ± 0.08g
Namwa 244.4 ± 3.6d 3418.7 ± 3.1b 68.8 ± 2.1e 111.7 ± 1.4k 241.7 ± 5.2c 25.8 ± 0.2f 0.4 ± 0.07d 4.8 ± 0.04c
Muisa Tia 151.4 ± 2.0i 3161.4 ± 6.7d 80.5 ± 2.6d 122.1 ± 2.8j 204.3 ± 2.8f 30.9 ± 0.2c 0.5 ± 0.04d 4.0 ± 0.04e
FC06-02 167.7 ± 1.9h 1870.9 ± 3.7q 78.1 ± 1.4d 205.9 ± 2.8f 209.7 ± 1.8f 28.9 ± 0.1d 0.7 ± 0.04c 2.9 ± 0.07g
Tiparot 350.4 ± 2.3a 2590.1 ± 5.8m 69.9 ± 4.2e 249.1 ± 4.9d 334.9 ± 3.5a 25.3 ± 0.2g 0.6 ± 0.05c 4.7 ± 0.08d
Plantains
D’Angola 163.3 ± 3.6h 2986.4 ± 4.1f 78.9 ± 2.8d 91.1 ± 1.4l 121.8 ± 3.1m 24.9 ± 0.1g 0.3 ± 0.03e 4.6 ± 0.03d
Curare Enano 130.8 ± 2.6j 2787.6 ± 6.6j 55.4 ± 3.1f 71.0 ± 1.4m 120.8 ± 3.4m 20.9 ± 0.2i 0.6 ± 0.03c 2.2 ± 0.03i
Terra S. N. 180.3 ± 3.8g 4192.7 ± 6.2a 61.6 ± 2.4f 122.4 ± 3.5j 142.3 ± 3.2k 29.1 ± 0.2d 0.5 ± 0.06d 4.6 ± 0.08d
Tipo Velhaca 149.3 ± 2.6i 2921.3 ± 7.1g 107.7 ± 1.7b 126.1 ± 2.1j 131.6 ± 3.9l 24.5 ± 0.2g 0.7 ± 0.04c 2.8 ± 0.04g
Terra A. B 139.6 ± 3.6j 2999.3 ± 6.1f 86.6 ± 1.8c 66.9 ± 2.1m 119.4 ± 1.2m 25.9 ± 01f 0.7 ± 0.03c 2.7 ± 0.03g
Samurá B 137.3 ± 3.2j 1685.4 ± 7.0r 63.8 ± 3.1f 56.6 ± 3.4n 98.8 ± 1.6n 24.9 ± 0.2g 0.5 ± 0.04d 1.9 ± 0.04j

*The same lower case letters do not differ by Scott and Knott test (5%).

6
C.V. Borges, et al.
Fig. 2. Two-dimensional projection (A) and scores (B) of phenolic compounds
and mineral contents and the antioxidant activity of pulp (stage 2) of the Musa
spp. genotypes.
= − 0.08; Zn: r = 0.1; K: r = 0.1; Ca: r = 0.02; Fe: 0.07) (PC1 + and
PC2). PC1 showed the highest correlation (0.7–0.9) with the P, Mg, Zn
and calcium contents. Regarding the genotypes, ‘Pacha Nadam’, ‘Khai’
and ‘Ouro da Mata’ displayed the highest correlation with PC1 (Fig. 2
and Table 3), highlighting these minerals.
Despite the abundance of literature about the contents of chemical
compounds in fruits and vegetables, the studies on the correlation be-
tween the mineral content and the antioxidant activity of these food
extracts are rare. When the minerals are bound to the phenolic com-
pounds, they can show considerable synergism in their antioxidant
capacity because certain minerals can act as electrons donors and have
their charges quickly stabilised by the polyphenolic structures
(Sant’Ana et al., 2012).
Sulaiman et al. (2011) did not find any correlation between the
7
Food Research International 132 (2020) 109061
mineral content and antioxidant activity of banana cultivars from Ma-
laysia, corroborating the findings described herein. In addition, an
imbalance of minerals can change the flavonoid content, a potent an-
tioxidant present in Musa spp. fruit. The deficiency of P can lead to an
increase in the flavonoid level, which can also help to explain, at least
partially, the negative correlation between the content of phenolic
compounds and the minerals found in the present study (Lillo, Lea, &
Ruoff, 2008; Tewari, Kumar, & Sharma, 2006).
3.3. Effect of the ripening stage on the contents of phenolic compounds in
pulps of Musa spp. fruit
The amounts of TP varied widely among the analysed genotypes
(74.64 to 506.12 mg GAE/100 g, Fig. 1). The ripening stage influenced
the TP content of the cultivars. In general, lower concentrations of TP
were found in bananas during stage 7, mainly in dessert ones, probably
because of the metabolic process associated with ripening. Throughout
fruit ripening, there is a marked decrease in the phenolic content, which
might be linked to the activity of oxidative enzymes, such as polyphenol
oxidase (Parr & Bolwell, 2000). In addition, in more advanced ripening
stages, a decrease in the quantity of these secondary metabolites might
result from the decrease of the enzymatic activity in their biosynthesis
pathways (i.e., phenylalanine ammonia-lyase EC 4.3.1.5) (Starrett &
Laties, 1991). Studies have verified that the phenolic content in plan-
tain pulps and peels increased until ripening stage 5, and decreasing in
stage 7, corroborating with the profile of many genotypes studied
(Fig. 1) (Vu et al., 2019). However, in some genotypes (e.g., ‘Ney
Poovan’ and ‘D'Angola’), there was an increase in these compounds in
stage 7, suggesting the need of further studies.
The pulp of the cooking genotypes ‘Tiparot’ and ‘Pelipita’ and the
dessert banana ‘Ney Poovan’ stood out for their phenolic amounts,
among all genotypes. ‘Pelipita’ presented a high TP content, mainly in
the green fruit (506.12 mg GAE/100 g) (Fig. 1) and ‘Tiparot’ showed
larger amounts of phenolics, both in green and in ripe fruit (stage 5,
658.80 mg GAE/100 g). ‘Ney Poovan’ was also noticeable for its in-
crease in these those secondary metabolites over the ripening stages
(460.19 mg GAE/100 g d.w.). In studies with ‘Kluai Leb Mue Nang’
bananas (AA group), augmented amounts of TP were found in more
advanced ripening stages (stage 5 and 7, with no difference between
them) (Youryon & Supapvanich, 2017).
Among the plantains, ‘Samurá B’ showed the highest TP content
(432.06 mg GAE/100 g d.w.) in mature fruits (stage 5). In green and
over-ripe fruit (yellow with black spots), genotype ‘D'Angola’ revealed
the highest amounts of these compounds (Fig. 1). Similar levels of
phenolic compounds described here were found by Tsamo et al. (2014)
in ‘Pelipita’ (319.5 mg GAE/100 g fresh weight, f.w.). However, in
‘Pelipita’ genotype, meaningful changes were detected over the ri-
pening stages, presenting 506.12 mg GAE/100 g d.w. (223.76 mg GAE
100 g f.w.) in stage 2, decreasing to 405.09 mg GAE/100 g d.w.
(158.47 mg GAE/100 g f.w.) in stage 5 and then slightly increasing to
432.28 mg GAE/100 g d.w. (162.62 mg GAE/100 g f.w.) in stage 7.
Such values are slightly below those reported by Tsamo et al. (2014).
Differences in the amounts observed between these studies are probably
due to the distinct extraction methods and analytical protocols used, as
well as the agroecological conditions (e.g., climate, soil, altitude) at the
respective cultivation sites.
3.4. Impact of the ripening stage on the chromatographic profile of phenolics
in pulps of Musa spp. fruit
The eight genotypes that stood out regarding their phenolic contents
were selected for further HPLC analysis. The chromatographic profiles
allowed to identify gallic acid, hydroxybenzoic acid and the flavonoids
catechin, epigallocatechin and quercetin (Table 5). The dessert banana
‘Ney Poovan’ and the cooking banana ‘Tiparot’ presented the highest
flavonoids and phenolic acids contents, regardless of the ripening stage.
C.V. Borges, et al. Food Research International 132 (2020) 109061

Table 5
Phenolic acids and flavonoid contents (µg/g d.w.) in pulps of Musa spp. genotypes determined by HPLC.
Genotypes Ripening stage Phenolic acids Flavonoids

Gallic acid Hydroxybenzoic acid Mean Epigallocatechin Catechin Quercetin Mean

Dessert banana
Ney Poovan 2 241.7 ± 10.7bB 39.9 ± 0.2eC 140.8 53.6 ± 0.5dC 262.9 ± 5.4bC 75.1 ± 9.5bB 130.5
5 254.4 ± 7.7aA 85.9 ± 3.3aA 170.2 60.8 ± 0.4dB 288,1 ± 1.8bA 98.1 ± 1.8bA 149.0
7 184.9 ± 7.7bC 68.9 ± 0.2aB 126.9 78.5 ± 0.3dA 269,4 ± 0.2aB 79.9 ± 0.3aB 142.6
Prata-Anã 2 183.5 ± 5.6cA 67.6 ± 0.5bA 125.5 19.2 ± 0.2eB 88.0 ± 0.1eB 68.4 ± 0.7cA 58.5
5 146.5 ± 0.8bB 58.2 ± 0.1bB 102.4 28.2 ± 1.4eA 106.5 ± 1.2eA 44.6 ± 0.2dB 59.8
7 145.1 ± 1.5cB 50.8 ± 0.4bC 97.9 15.2 ± 0.3gC 51,1 ± 1.7fC 18.9 ± 0.1cC 28.4
Grande Naine 2 101.2 ± 0.9eA 53.5 ± 0.5dA 77.4 19.2 ± 0.3eA 88.5 ± 1.3eA 13.1 ± 0.2fA 40.3
5 86.6 ± 0.1dB 44.1 ± 0.3dB 65.4 19.8 ± 0.3fA 78.4 ± 0.4fB 10.0 ± 0.2gA 36.1
7 86.7 ± 0.4eB 34.3 ± 0.6dC 60.5 11.2 ± 1.9hB 23.3 ± 1.8gC 6.9 ± 1.2dB 13.8
Cooking banana
Tiparot 2 352.0 ± 4.4aA 28.5 ± 0.2fB 190.3 121.0 ± 4.8bB 459.1 ± 2.2aB 142.0 ± 0.3aB 240.7
5 249.2 ± 4.5aB 58.2 ± 0.3bA 153,7 149.8 ± 0.3bA 641,6 ± 2.9aA 153.0 ± 1.4aA 314.8
7 193.9 ± 1.2aC 28.2 ± 0.1eB 111.1 119.5 ± 0.4bB 117.2 ± 0.7cC 32.9 ± 0.4bC 89.9
Pelipita 2 70.5 ± 6.4f C 14.1 ± 0.6gC 42.3 14.8 ± 3.7fB 118.3 ± 3.5dB 35.3 ± 0.6dA 56.1
5 80.7 ± 1.2dA 43.3 ± 1.4dA 62.0 28.9 ± 4.2eA 122.1 ± 3.5dB 28.2 ± 1.7eB 60.0
7 67.8 ± 0.8fC 29.1 ± 0.2eB 48.5 31.4 ± 0.2fA 117.1 ± 2.3cA 17.6 ± 0.3cC 55.4
Plantain
D’Angola 2 151.6 ± 1.2dA 72.7 ± 0.2aA 112.2 77.5 ± 2.8cA 82.8 ± 2.3fA 20.1 ± 0.5eA 60.1
5 95.6 ± 4.1cB 56.9 ± 2.7bB 76.3 57.7 ± 1.9dB 70.6 ± 1.7gB 16.1 ± 0.2gB 48.1
7 79.7 ± 4.5eC 38.6 ± 2.4cC 59.2 59.6 ± 1.8eB 52.9 ± 0.2fC 23.2 ± 0.2cA 45.2
Terra S. N. 2 159.9 ± 1.5dB 12.7 ± 0.4gB 86.3 51.2 ± 0.9dC 24.0 ± 1.1gC 66.4 ± 2.3cA 47.2
5 248.5 ± 4.5aA 16.7 ± 0.4eA 132.6 134.5 ± 6.2cB 42.6 ± 0.3hB 26.4 ± 0.2eB 67.8
7 146.5 ± 7.7cC 10.7 ± 0.2fB 78.6 170.3 ± 4.7aA 79.9 ± 2.1eA 26.2 ± 2.7bB 92.1
Samurá B 2 154.9 ± 1.3dA 72.3 ± 0.6aA 113.6 147.9 ± 1.5aB 164.2 ± 6.5cA 65.1 ± 6.4cB 125.7
5 142.0 ± 5.2bB 54.7 ± 0.2cB 98.4 158.5 ± 0.3aA 160.5 ± 0.2cA 82.1 ± 2.3cA 133.7
7 123.6 ± 13.4dC 37.5 ± 0.2cC 80.5 105.7 ± 0.2cC 107.2 ± 1.8dB 31.6 ± 2.8bC 81.5

*The same lower case letters (genotypes) and uppercase letters (ripening stages) do not differ by Scott and Knott test (5%) and Tukey test (5%), respectively.

Among the plantains, the genotype ‘Samurá B’ presented the highest
levels.
In a tentative attempt to establish a descriptive model of grouping
samples based on the fruit ripening stages as a function of their contents
of TP, the PCA was applied to the chemical data set. PC1 and PC2 ac-
counted for 80.52% of the data variance (Fig. 3). PC1 explained 57.81%
of the data variance and was responsible for separating the genotypes
with higher TP content (PC1 + ) from those with less TP (PC1-). In-
terestingly, ‘Ney Poovan’, ‘Tiparot’ and ‘Samurá B’ were grouped be-
cause of their higher amounts of phenolic acids and flavonoids, re-
gardless of their ripening stage (Fig. 3). PC2 explained 22.7% of the
data variance and was responsible for separating the genotypes by their
phenolic profiles. The dessert banana ‘Ney Poovan’ (stage 5 and 7), the
plantain ‘Samurá B’ (stage 2) and the cooking banana ‘Tiparot’ (stage 5)
presented the highest contents of hydroxybenzoic acid, grouping in
PC2+. The genotype ‘Tiparot’ also presented the highest content of
catechin, quercetin and gallic acid (stage 2 and 5). Superior amounts of
epigallocatechin were found in more advanced stages of fruit ripening
for the plantains ‘Samurá B’ (stage 5) and ‘Terra Sem Nome’ (stage 7),
which were grouped in PC2- (Fig. 3).
There was no discernible genotype grouping regarding the ripening
stages as a function of their phenolic compound amounts (Fig. 3).
However, the most values of the phenolic acid and flavonoid contents
increased until stage 5 (ripe fruit), decreasing in more advanced ri-
pening stages (stage 7) (Table 5). Some authors reported increases of
these compounds during fruit ripening and a decrease in over-ripe
stages (Campuzano, Rosell, & Cornejo, 2018; Youryon & Supapvanich,
2017), consistent with the results found in the present study. In addi-
tion, fruit with more light-coloured pulps tend to present higher fla-
vonoids contents, differently from what was found in other antioxidant
compounds (i.e., carotenoids, pro-vitamins) in Musa spp. (Borges et al.,
2018).
Pearson’s correlation analysis revealed that flavonoids were the
compounds most strongly correlated with the antioxidant activity.
Catechin and quercetin presented the highest correlation values,
regardless of the antioxidant method used (FRAP: r = 0.92 and 0.85;
DPPH: r = 0.85 and 0.80; ABTS: r = 0.78 and 0.77, p < 0.05, re-
spectively). Among the phenolic acids, gallic acid presented a positive
and significant correlation with the antioxidant activity (DPPH:
r = 0.70; ABTS: r = 0.67; FRAP: r = 0.65, p < 0.05), but lower than
that detected for flavonoids (e.g., catechin and quercetin).
Phenolic compounds are considered the most important anti-
oxidants present in plants. The antioxidant activity of these compounds
depends on the type and quantity of phenolic compounds in the plant
cell (Anyasi et al., 2018; Demiray, Pintado, & Castro, 2009), as noted in
the present study. Thus, the phenolic profile of each sample is essential
to elucidate which compound is predominantly responsible for the
antioxidant activity observed. Besides, the interactive effect of those
secondary metabolites should be considered when determining the
antioxidant action found in plant extracts. Our results show, for ex-
ample, that hydroxybenzoic acid (FRAP: r = 0.07; DPPH: r = 0.16;
ABTS: r = 0.24, p < 0.05) and epigallocatechin (FRAP: r = 0.32;
FRAP: r = 0.22; ABTS: r = 0.15, p < 0.05) presented low correlation
with the antioxidant activity in Musa spp. fruits, differing from other
analysed phenolic compounds, such as catechin and quercetin, men-
tioned above.
3.5. Impact of the thermal processing on the phenolic compounds content in
cooking banana and plantains
To establish a descriptive model of the grouping of genotypes ac-
cording to their phenolic compounds and thermal processing, the
quantitative data of the secondary metabolites resulting from the HPLC
analysis were applied to PCA. The PC2 axis covered 14.82% of the total
data variance and was responsible for separating the genotypes by their
phenolic contents, such as noted for the ‘D'Angola’ plantain (PC2 + ),
which had the highest epigallocatechin content (Fig. 4 and Table 6).
However, when the fruit of this genotype were submitted to the frying
treatment, a significant decrease in that flavonoid was detected. The
cooking banana ‘Pelipita’ (PC2+) showed the highest contents of the

8
C.V. Borges, et al.
Fig. 3. Two-dimensional projection (A) and scores (B) of phenolic compounds
(gallic acid, hydroxybenzoic acid, epigallocatechin, catechin, and quercetin) of
the eight genotypes of bananas and plantains evaluated during the ripening
stages (stages 2, 5 and 7).
flavonoids catechin and quercetin, as well as of hydroxybenzoic acid,
regardless of the thermal treatment adopted.
For the cooking methods investigated, the pulps of non-peeled fruits
displayed significant increases in the phenolic compounds, mainly the
phenolic acids (Table 6). This result can be explained by the high
amount of these secondary metabolites in the peels, which might have
migrated to the pulp (Tsamo et al., 2015). Additionally, certain geno-
type-dependent discrepancies in the contents of phenolics were found
after exposing the fruit biomass to the thermal treatments. For instance,
the cooking banana ‘Pelipita’ presented, in general, an increase in the
amounts of phenolic acids and flavonoids (epigallocatechin and quer-
cetin) after the thermal processing when compared with the plantain
‘D'Angola’ (Table 6). In general, the cooking bananas were less firm
than the plantains, which are characterised by a high fruit firmness,
even in more advanced ripening stages (Borges et al., 2019). Accord-
ingly, the cooking banana ‘Pelipita’ was less firm than the plantain
‘D'Angola’ at the moment of the thermal processing, allowing better
extraction of these compounds in cooking bananas.
9
Food Research International 132 (2020) 109061
Fig. 4. Two-dimensional projections (A) and scores (B) of phenolic compounds
(gallic acid, hydroxybenzoic acid, epigallocatechin, and quercetin) of the
plantain 'D'Angola' (D) and the cooking banana 'Pelipita' (P), submitted to the
different types of thermal processing.
4. Conclusion
Important differences in the primary and secondary metabolites
composition of the investigated Musa spp. genotypes were found.
Plantains and cooking bananas presented a high content of resistant
starch. The lowest amounts of phenolic compounds were found in the
plantains. Contrarily, these secondary metabolites were detected in the
highest levels in the dessert and cooking bananas, which also revealed
superior amounts of mineral nutrients, as found in the genotypes ‘Khai’
and ‘Ouro da Mata’ (dessert banana) and ‘Pacha Nadam’ (cooking ba-
nana). Our results also indicate that the fruit peel has levels of phenolic
compounds and minerals superior to the pulp (e.g.,), being an im-
portant by-product source of bioactive compounds of interest of the
pharmaceutical and food industries.
Catechin and quercetin were the compounds that contributed
mostly to the antioxidant activity of the Musa spp. germplasm. The
dessert banana ‘Ney Poovan’ and the cooking banana ‘Tiparot’ pre-
sented the highest content of these compounds. Besides, the content of
phenolic compounds was affected by the ripening stage, and superior
values were found in the ripe fruit (stage 5 - yellow with green). In
C.V. Borges, et al. Food Research International 132 (2020) 109061

Table 6
Phenolic and flavonoid contents (µg/g d.w.) after pulp cooking processes of the ‘D’Angola’ plantain and ‘Pelipita’ cooking banana measured by HPLC.
Cooking processes Phenolic acids Flavonoids

Gallic acid Hydroxybenzoic acid Mean Epigallocatechin Catechin Quercetin Mean

Pelipita
Boiling with peel 160.6 ± 7.2a 97.8 ± 1.4a 129.2 55.9 ± 0.4a 83.5 ± 0.9b 60.0 ± 0.3a 66.5
Boiling without peel 115.7 ± 0.6b 67.2 ± 0.2c 91.5 36.8 ± 0.3c 79.9 ± 0.3c 49.6 ± 0.2b 55.4
Microwave with peel 91.7 ± 4.6c 93.1 ± 0.5b 92.4 48.7 ± 1.4b 89.3 ± 2.3a 50.4 ± 0.9b 62.8
Microwave without peel 61.6 ± 1.5e 47.8 ± 0.5d 54.7 25.9 ± 0.2d 64.5 ± 0.7d 16.4 ± 0.2d 35.6
Stir-frying 33.8 ± 1.8f 25.2 ± 0.2f 29.5 1.4 ± 0.1e 13.8 ± 0.1e 7.9 ± 0.3e 7.7
Raw 75.5 ± 0.1d 43.9 ± 0.5e 59.7 27.8 ± 1.0d 79.8 ± 0.1c 38.9 ± 0.2c 48.8
D’Angola
Boiling with peel 97.0 ± 0.2a 60.1 ± 0.2a 78.6 92.1 ± 0.8a 59.4 ± 0.1a 30.5 ± 0.3a 60.7
Boiling without peel 84.6 ± 0.1c 40.9 ± 0.1c 62,8 73.3 ± 0.6c 51.9 ± 0.7b 28.2 ± 0.1b 51.1
Microwave with peel 92.5 ± 0.1b 46.1 ± 1.0b 69.3 80.0 ± 1.9b 60.6 ± 0.1a 29.9 ± 0.2a 56.8
Microwave without peel 47.7 ± 0.1e 30.0 ± 0.1d 38.9 55.2 ± 0.3d 52.2 ± 0.6b 16.3 ± 0.1d 41.2
Stir-frying 17.3 ± 0.2f 10.5 ± 0.2e 13.9 3.2 ± 0.4e 15.1 ± 0.2d 4.4 ± 0.4e 7.6
Raw 76.8 ± 0.2d 45.9 ± 0.2b 61.4 56.8 ± 0.6d 41.9 ± 0.6c 27.3 ± 0.4c 42.0

*The same lower case letters do not differ by Tukey test (5%).

addition, the thermal process increased these secondary metabolites of
the Musa spp. fruit, mainly the ebullition cooking method (fruit with
peel), which should be the preferred method in domestic preparation,
regardless of the cultivar.
Acknowledgements
The authors gratefully acknowledge the financial support of Sao
Paulo Research Foundation (FAPESP) for financial support of this work
(grant number 2016/22665-2) and for scholarship provided to C.V.
Borges (grant number 2016/00972-0) and to M.A.F. Belin (grant
number 2017/23488-0) and National Council for Scientific and
Technological Development (CNPq, Brazil) (grant number 305177/
2015-0). The researcher fellowship from CNPq (grant number 307099/
2015-6) on behalf of Marcelo Maraschin is acknowledged.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.foodres.2020.109061.
References
Altundag, H., & Tuzen, M. (2011). Comparison of dry, wet and microwave digestion
methods for the multi element determination in some dried fruit samples by ICP-OES.
Food and Chemical Toxicology, 49(11), 2800–2807.
Anyasi, T. A., Jideani, A. I. O., & Mchau, G. R. A. (2018). Phenolics and essential mineral
profile of organic acid pretreated unripe banana flour. Food Research International,
104(May 2017), 100–109.
Barrameda-Medina, Y., Blasco, B., Lentini, M., Esposito, S., Baenas, N., Moreno, D. A., &
Ruiz, J. M. (2017). Zinc biofortification improves phytochemicals and amino-acidic
profile in Brassica oleracea cv. Bronco. Plant Science, 258, 45–51.
Benzie, I. F. F., & Strain, J. J. (1996). The Ferric Reducing Ability of Plasma (FRAP) as a
Measure of “Antioxidant Power”: The FRAP Assay. Analytical Biochemistry, 239(1),
70–76.
Bergmeyer, H. U., & Bernet, E. (1974). Determination of glucose with glucose oxidase and
peroxidase. In H. U. Bergmeyer (Ed.). Methods of enzymatic analysis (pp. 1204–1212).
New York: Academic Press.
Bi, Y., Zhang, Y., Gu, Z., Cheng, L., Li, Z., Li, C., & Hong, Y. (2019). Effect of ripening on in
vitro digestibility and structural characteristics of plantain (Musa ABB) starch. Food
Hydrocolloids, 93, 235–241.
Borges, C. V., Amorim, V. B. O., Ramlov, F., Ledo, C. A. S., Donato, M., Maraschin, M., &
Amorim, E. P. (2014). Characterisation of metabolic profile of banana genotypes,
aiming at biofortified Musa spp. cultivars. Food Chemistry, 145, 496–504.
Borges, C. V., Amorim, E. P., Leonel, M., Gomez, H. A. G., Santos, T. P. R. dos, Ledo, C. A.
da S., …& Lima, G. P. P. (2019). Post-harvest physicochemical profile and bioactive
compounds of 19 bananas and plantains genotypes. Bragantia, 78(2), 284–296.
Borges, C. V., Belin, M. A. F., Amorim, E. P., Minatel, I. O., Monteiro, G. C., Gomes, H. A.
G., ... Lima, G. P. P. (2019). Bioactive amines changes during the ripening and
thermal processes of bananas and plantains. Food Chemistry, 298(June), 125020.
Borges, C. V., Minatel, I. O., Amorim, E. P., Belin, M. A. F., Gomez-Gomez, H. A., Correa,
C. R., & Lima, G. P. P. (2018). Ripening and cooking processes influence the car-
otenoid content in bananas and plantains (Musa spp.). Food Research International,
124, 129–136.
Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free radical method to
evaluate antioxidant activity. LWT - Food Science and Technology, 28(1), 25–30.
Campuzano, A., Rosell, C. M., & Cornejo, F. (2018). Physicochemical and nutritional
characteristics of banana flour during ripening. Food Chemistry, 256(November
2017), 11–17.
Demiray, S., Pintado, M. E., & Castro, P. M. L. (2009). Evaluation of phenolic profiles and
antioxidant activities of Turkish medicinal plants: Tilia argentea. Crataegi folium
leaves and Polygonum bistorta roots, 3(6), 74–79.
FAO. (2017). Food and Agriculture Organization. Production crop data. Food and
Agriculture Organization. Production Crop Data. (Available in: < http://www.fao.
org/faostat/en/#data > . Accessed in: 02 February, 2019.
Genc, Y., Humphries, J. M., Lyons, G. H., & Graham, R. D. (2005). Exploiting genotypic
variation in plant nutrient accumulation to alleviate micronutrient deficiency in
populations. Journal of Trace Elements in Medicine and Biology, 18(4), 319–324.
Ghag, S., & Ganapathi, T. (2018). Banana and Plantains: Improvement, Nutrition, and
Health. In: K. G. R. J.-M. Mérillon (Ed.), Bioactive Molecules in Food (pp. 1–20).
Springer International Publishing AG.
Goñi, I., Garcia-Alonso, A., & Saura-Calixto, F. (1997). A starch hydrolysis procedure to
estimate glycemic index. Nutrition Research, 17(3), 427–437.
Goñi, I., García-Diz, L., Mañas, E., & Saura-Calixto, F. (1996). Analysis of resistant starch:
A method for foods and food products. Food Chemistry, 56(4), 445–449.
Institute of Medicine (IOM). (2000). Dietary Reference Intakes for Vitamin C, Vitamin E,
Selenium, and Carotenoids. (D. National Academy Press, Washington, Ed.), National
Academy Press, Washington, DC.
Khoozani, A. A., Birch, J., & Bekhit, A. E. A. (2019). Production, application and health
effects of banana pulp and peel flour in the food industry. Journal of Food Science and
Technology, 56, 548–559.
Lillo, C., Lea, U. S., & Ruoff, P. (2008). Nutrient depletion as a key factor for manipulating
gene expression and product formation in different branches of the flavonoid
pathway. Plant, Cell and Environment, 31, 587–601.
Parr, A. J., & Bolwell, G. P. (2000). Phenols in the plant and in man. The potential for
possible nutritional enhancement of the diet by modifying the phenols content or
profile. Journal of the Science of Food and Agriculture, 80, 985–1012.
Pérez-Jiménez, J., & Saura-Calixto, F. (2018). Fruit peels as sources of non-extractable
polyphenols or macromolecular antioxidants: Analysis and nutritional implications.
Food Research International, 111(May), 148–152.
Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., & Rice-Evans, C. (1999).
Antioxidant Activity Applying an Improved Abts Radical cation decolorization assay.
Free Radical Biology & Medicine, 26(98), 1231–1237.
Sant’Ana, L., Sousa, J. P. L. M., Salgueiro, F. B., Lorenzon, M. C. A., & Castro, R. (2012).
Characterization of monofloral honeys with multivariate analysis of their chemical
profile and antioxidant activity. Journal of Food Science, 71(1), 135–140.
Sardá, F. A. H., Giuntini, E. B., Gomez, M. L. P. A., Lui, M. C. Y., Negrini, J. A. E., Tadini,
C. C., ... Menezes, E. W. (2016). Impact of resistant starch from unripe banana flour
on hunger, satiety, and glucose homeostasis in healthy volunteers. Journal of
Functional Foods, 24, 63–74.
Singleton, V. L., & Rossi, J. A... J. (1965). Colorimetry of Total Phenolics with
Phosphomolybdic-Phosphotungstic Acid Reagents. Am. J. Enol. Vitic., 16(3),
144–158. Retrieved from http://www.ajevonline.org/content/16/3/144.abstract.
Soltani, M., Alimardani, R., & Omid, M. (2011). Evaluating banana ripening status from
measuring dielectric properties. Journal of Food Engineering, 105(4), 625–631.
Starrett, D. A., & Laties, G. G. (1991). Involvement of wound and climacteric ethylene in
ripening avocado discs. Plant Physiology, 97(2), 720–729.
Sulaiman, S. F., Yusoff, N. A. M., Eldeen, I. M., Seow, E. M., Sajak, A. A. B., Supriatno, &
Ooi, K. L. (2011). Correlation between total phenolic and mineral contents with
antioxidant activity of eight Malaysian bananas (Musa sp.). Journal of Food
Composition and Analysis, 24(1), 1–10.
Tewari, R. K., Kumar, P., & Sharma, P. N. (2006). Magnesium deficiency induced oxi-
dative stress and antioxidant responses in mulberry plants. Scientia Horticulturae, 108,

10
C.V. Borges, et al.
7–14.
Tsamo, C. V. P., Andre, C. M., Ritter, C., Tomekpe, K., Newilah, G. N., Rogez, H., &
Larondelle, Y. (2014). Characterization of Musa sp. fruits and plantain banana ri-
pening stages according to their physicochemical attributes. Journal of Agricultural
and Food Chemistry, 62(34), 8705–8715.
Tsamo, C. V. P., Herent, M. F., Tomekpe, K., Emaga Quetin-Leclercq, T. H., Rogez, H.,
Larondelle, Y., & Andre, C. (2015). Phenolic profiling in the pulp and peel of nine
plantain cultivars (Musa sp.). Food Chemistry, 167, 197–204.
Vu, H. T., Scarlett, C. J., & Vuong, Q. V. (2019). Changes of phytochemicals and
11
Food Research International 132 (2020) 109061
antioxidant capacity of banana peel during the ripening process; with and without
ethylene treatment. Scientia Horticulturae, 253(January), 255–262.
Wall, M. M. (2006). Ascorbic acid, vitamin A, and mineral composition of banana (Musa
sp.) and papaya (Carica papaya) cultivars grown in Hawaii. Journal of Food
Composition and Analysis, 19(5), 434–445.
Youryon, P., & Supapvanich, S. (2017). Physicochemical quality and antioxidant changes
in ‘Leb Mue Nang’ banana fruit during ripening. Agriculture and Natural Resources,
51(1), 47–52.