|

Received: 19 February 2017    Accepted: 31 August 2017

DOI: 10.1111/eva.12549

ORIGINAL ARTICLE

Long-­term on-­farm participatory maize breeding by stratified

mass selection retains molecular diversity while improving
agronomic performance

Mara Lisa Alves1  | Maria Belo1 | Bruna Carbas2 | Cláudia Brites3 | Manuel Paulo3 | 

Pedro Mendes-Moreira3 | Carla Brites2 | Maria do Rosário Bronze1,4,5 | 
Zlatko Šatović6 | Maria Carlota Vaz Patto1

1
Instituto de Tecnologia Química e Biológica
António Xavier, Universidade Nova de Lisboa, Abstract
Oeiras, Portugal Modern maize breeding programs gave rise to genetically uniform varieties that can
2
Instituto Nacional de Investigação Agrária e
affect maize’s capacity to cope with increasing climate unpredictability. Maize popula-
Veterinária, Oeiras, Portugal
3 tions, genetically more heterogeneous, can evolve and better adapt to a broader range
Departamento de Ciências Agronómicas, 
Escola Superior Agrária de Coimbra, Coimbra, of edaphic–climatic conditions. These populations usually suffer from low yields; it is
Portugal
therefore desirable to improve their agronomic performance while maintaining their
4
Faculdade de Farmácia, Universidade de
Lisboa, Lisboa, Portugal valuable diversity levels. With this objective, a long-­term participatory breeding/on-­
5
Instituto de Biologia Experimental e farm conservation program was established in Portugal. In this program, maize popula-
Tecnológica, Oeiras, Portugal tions were subject to stratified mass selection. This work aimed to estimate the effect
6
Faculty of Agriculture, Department of Seed
of on-­farm stratified mass selection on the agronomic performance, quality, and mo-
Science and Technology, University of Zagreb,
Zagreb, Croatia lecular diversity of two historical maize populations. Multilocation field trials, compar-
ing the initial populations with the derived selection cycles, showed that this selection
Correspondence
Mara Lisa Alves, Instituto de Tecnologia methodology led to agronomic improvement for one of the populations. The molecu-
Química e Biológica António Xavier,
lar diversity analysis, using microsatellites, revealed that overall genetic diversity in
Universidade Nova de Lisboa, Oeiras,
Portugal. both populations was maintained throughout selection. The comparison of quality pa-
Email: [email protected]
rameters between the initial populations and the derived selection cycles was made
Funding information using kernel from a common-­garden experiment. This analysis showed that the major-
European Union’s 7th Framework Programme,
ity of the quality traits evaluated progressed erratically over time. In conclusion, this
Grant/Award Number: FP7/2007-2013;
Horizon 2020 Framework Programme, breeding approach, through simple and low-­cost methodologies, proved to be an alter-
Grant/Award Number: 245058 and
native strategy for genetic resources’ on-­farm conservation.
633571; Fundação para a Ciência e a
Tecnologia, Grant/Award Number: UID/
Multi/04551/2013, SFRH/BD/77723/2011 KEYWORDS
and IF/01337/2014 ear traits, microsatellites, molecular diversity, on-farm conservation, open-pollinated
populations, participatory plant breeding, yield, Zea mays L

1 |  INTRODUCTION and therefore on food production is exacerbated by greater crop

uniformity (Ceccarelli et al., 2010). An increasing number of studies
Climate change represents a challenge to food security (Wheeler & von show that biodiversity improves the capacity of agroecosystems to
Braun, 2013). The negative impact of climate change on agriculture cope with extreme weather events and climate variability (Khoury

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2017 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd

Evolutionary Applications. 2017;1–17. wileyonlinelibrary.com/journal/eva |  1

  
 |
2       ALVES et al.

et al., 2014; Ortiz, 2011), allowing crops’ evolution and adaptation to been associated with the prevention of modern lifestyle-­related de-
specific edaphic–climatic conditions (Ceccarelli, 2015). This is partic- generative disease (Liu, 2003). In that regard, maize displays a con-
ularly important in the context of low-­input/organic production sys- siderable natural variation for content and composition of antioxidant
tems, more prone to biotic and abiotic constrains and in which crop compounds such as carotenoids (Owens et al., 2014) and tocopherols
resilience is fundamental. The greater uniformity of crops is specifi- (Lipka et al., 2013). However, little is known about the phytochemical
cally a concern for maize, wheat, and rice, which alone provide 60% profiles, antioxidant activity, or organoleptic quality of the different
of the calories in the human diet. In these three crops, recent plant Portuguese maize open-­pollinated populations with high technologi-
breeding has led to extreme genetic uniformity (Ceccarelli, Galie, & cal ability for bread production.
Grando, 2013). As reviewed by Hellin, Bellon, and Hearne (2014), it is With the development of modern sustainable low-­input agricul-
important that plant breeding reach a compromise by developing not ture in industrialized countries, for economic and environmental rea-
only higher-­yielding but also stress-­tolerant cultivars, to allow them sons, emphasis has been placed on local adaptation, on preservation
to cope and adapt when faced with different environmental condi- of genetic diversity, and on quality (Cleveland, Soleri, & Smith, 1999).
tions. In the case of maize, the more heterogeneous open-­pollinated Conventional plant breeding has been successful in favorable environ-
populations, adapted to specific environmental conditions and human ments, but is less successful in traditional low-­input or organic farming
uses, have progressively been replaced in the last century by homo- systems with higher stress growing conditions, especially in small-­scale
geneous, higher-­yielding commercial hybrids (Pingali, 2001). Still, farms (Vaz Patto et al., 2013). Under this scenario, participatory plant
open-­pollinated populations cultivation has been maintained, often in breeding (PPB) programs are arising worldwide to meet the needs of
marginal lands or low-­input systems where commercial hybrids are not farmers in low-­input and organic environments that are normally over-
well adapted (Vaz Patto et al., 2013). They may also be kept by their looked by conventional crop breeders (Vaz Patto et al., 2013).
dietary or nutritional value, taste, or for the price premium they attract Participatory plant breeding differs from conventional breeding
because of high-­quality traditional properties that compensate for mainly because of the active participation of other actors apart from
lower yields (Jarvis, Hodgkin, Sthapit, Fadda, & Lopez-­Noriega, 2011). breeders, such as farmers and/or consumers, in the breeding program.
Portugal was one of the first European countries to adopt maize Those actors can assume an active role in the establishment of the
and one of the few where historical maize populations can still be breeding objectives and influence or actively participate in the breed-
found under cultivation (Vaz Patto et al., 2013). The resilience of these ing activities. In the case of on-­farm participatory breeding, the selec-
maize populations in the Portuguese scenario can be partially explained tion is made at the farmer’s field, in a partnership between breeder
by their technological quality in maize bread production (Vaz Patto and farmer, with the farmer establishing the breeding objectives (Vaz
et al., 2013). The Portuguese ethnic maize-­based bread, named broa, Patto et al., 2013). Taking into consideration the central role attributed
is highly accepted for its distinctive sensory characteristics (Carbas to farmers on this breeding approach, their acceptance and enthusi-
et al., 2016). This bread is traditionally manufactured using local maize asm while participating in the program has been identified as one of
populations and still plays an important economic and social role on the key aspects for the success of on-­farm participatory plant breed-
Central and Northern rural communities of the country (Vaz Patto, ing (Vaz Patto et al., 2013). This type of decentralized PPB improves
Moreira, Carvalho, & Pego, 2007). Broa is traditionally made with more breeding efficiency as it increases the ratio of the number of varieties
than 50% maize flour mixed with rye and/or wheat flour by a mainly adopted by farmers, as it is the farmer’s choice to adopt those variet-
empirical process (Brites, Trigo, Santos, Collar, & Rosell, 2010). This ies into the program; it also increases traits’ response to selection, as
process normally involves the mixing of the sieved wholemeal maize selection is being made in the targeted environment (Ceccarelli, 2015).
flour, with hot water, rye and/or wheat flour (in a variable proportion), In 2012, Ceccarelli, Al-­Yassin, Goldringer, Mendes-­Moreira, and
and yeast from leavened dough from late broa, acting as sourdough Chable (2012) published the results of a survey on the previous major
(Brites et al., 2010). PPB experiences worldwide. Of the 22 active PPB programs presented
In what concerns broa bread quality, differences between the in that report, three are in maize and are located in Portugal, China, and
higher-­yielding dent hybrids and the hard endosperm Portuguese Nepal. The Portuguese participatory maize breeding program started
open-­pollinated populations have been recently determined (Carbas in 1984 and initially had as its main objective the improvement of the
et al., 2016). In that work, it was shown that the broa produced with agronomic performance of historical maize populations, functioning in
the hybrid dent varieties had higher specific volume. However, sen- parallel as a strategy for the on-­farm conservation of those plant ge-
sory analysis showed a preference for the maize bread made using netic resources (Vaz Patto et al., 2013).
Portuguese open-­pollinated populations due to better mouthfeel The methodologies implemented in every breeding program are
flavor and texture (Carbas et al., 2016). Parameters associated with dependent on the type of reproductive system of the crop. In natu-
aroma or flavor (e.g., volatile aldehydes; Klensporf & Jelén, 2005, rally cross-­pollinated species, such as maize, improvement of open-­
and texture (e.g., viscosity parameters; Brites et al., 2010) can be im- pollinated populations can be achieved by recurrent mass selection if
portant in assessing the product’s quality and therefore need to be the pollinations are controlled and/or by the use of stratified selection
investigated. Additionally, bread nutritional value is another quality (Gardner, 1961). In the on-­farm breeding activities of the Portuguese
aspect with great importance. In recent years, consumption of par- maize participatory breeding program, as controlled pollinations are
ticular foods and food products, rich in antioxidant compounds, has time-­consuming, the use of stratified mass selection has been the
ALVES et al. |
      3

selected methodology. In mass selection, a fraction of individuals are studies can serve as a decision-­making tool for genetic resources’ man-
visually selected to form the following generation. As for stratified agement. This approach was applied, for example, in Lançon et al. (2008),
mass selection, prior to the selection of individuals (mass selection) in which the authors used molecular markers to access the genetic vari-
the field is first divided into smaller selection units (field stratification), ability of cotton populations of a participatory breeding program and,
minimizing the bias due to field heterogeneity. The differences among as a direct result, farmers were advised to use another breeding meth-
plants within field’s sections are more likely to be due to genetic dif- odology to increase the efficiency of the selection. Molecular markers
ferences than to environmental effects (Hallauer, Carena, & Miranda have been also used to assess the temporal variation on maize genetic
Filho, 2010). Stratified mass selection has been shown in the past to diversity due to human selection (Labate, Lamkey, Lee, & Woodman,
be a useful methodology for improving several agronomic traits in 1999 and Solomon, Martin, & Zeppa, 2010). In Labate et al. (1999),
maize, for example, for adapting exotic germplasm into breeding pro- the authors used molecular markers to study the effects of reciprocal
grams and target environments (Hallauer, 1999) or for yield improve- recurrent selection on Iowa Stiff Stalk Synthetic and Iowa Corn Borer
ment of open-­pollinated maize populations (Mendes-­Moreira, Pego, Synthetic maize populations, reporting a decrease in 39% of the mean
Vaz Patto, & Hallauer, 2008; Mendes-­Moreira et al., 2009; Smith, expected heterozygosity after 12 cycles of selection. In Solomon et al.
Castillo, & Gómez, 2001). (2010), using other types of molecular markers to study the effects of
In the Portuguese maize participatory breeding program, breed- reciprocal recurrent selection on tropical maize breeding populations,
ing activities were intended to occur mainly at the farmer’s field, with the authors reported a loss of 33%–37% of the alleles detected initially
breeder and farmer working side by side. Firstly, the selection method- after 11 cycles of selection. Nevertheless, as reviewed by Rauf, Teixeira
ologies were demonstrated by the breeder at each farmer’s field, and da Silva, Khan, and Naveed (2010), different plant breeding methods
afterward, the farmer conducted the same selection methodologies have shown different impacts on plant genetic diversity.
in the other part of the field. In this way, the farmer had a permanent In the specific case of the Portuguese PPB program, the agronomic
possibility to compare the effectiveness of the breeder’s advices and evaluation of the impact of breeding activities has only been per-
the breeder needed to respect the farmer’s management system (e.g., formed in two of the several maize populations integrated in the pro-
low-­input), advising only simple and low-­cost selection methodolo- gram (Mendes-­Moreira et al., 2008, 2009), and the temporal changes
gies based on population genetics theory, with the farmer keeping the on genetic diversity were only evaluated in one of those populations
decision power over the direction of selection. Besides the specific (Vaz Patto, Moreira, Almeida, Satovic, & Pego, 2008). Moreover, none
breeding objectives defined by each farmer for each maize popula- of these studies took into consideration quality aspects that should be
tion, in this program the farmer is advised by the breeder to select addressed in future breeding programs as the quality of these genetic
in the field by detasseling the undesirable plants before pollination resources for maize bread production seems to be a decisive aspect
(weakest and all that do not fit the desired ideotype, such as the pest for the on-­farm maintenance of the historical populations developed
and disease susceptible looking ones); the farmer is also advised to (Brites et al., 2010; Vaz Patto et al., 2013).
evaluate a few days before harvest the root and stalk quality by foot-­ The evaluation of the effect of stratified mass selection in the
kicking the plants at their base (at the first visible internodes). This also Portuguese maize participatory breeding program is crucial to under-
serves as an indirect measurement of pest tolerance, as the plant that stand whether the methodologies implemented in this program are
does not resist the impact and breaks down is eliminated. Additionally, effective or need to be revised in order to accomplish the defined
the farmer is advised to favor the selection of more prolific plants or breeding objectives. As a bulk of the harvested seed from each selec-
the ones with a lower ear insertion if that trait is among the farmer tion cycle was saved and kept in cold storage, it is possible to assess
desired ideotype. Prior to this selection, the field is first divided into the evolution of the maize populations within the breeding program.
smaller selection units (field stratification). After harvesting, a second Taking these factors into account, this work aimed to (i) evaluate
selection (postharvest) is conducted in the ears. This selection includes whether on-­farm stratified mass selection, in the context of long-­term
the specific breeding objectives of each population and the elimina- participatory research, was able to improve the agronomic perfor-
tion of unhealthy damaged ears. Selected ears are then shelled and mance of two historical maize open-­pollinated populations, Amiúdo
mixed together to form the next-­year generation. With this scheme, and Castro Verde, (ii) evaluate the effect of stratified mass selection
the selection pressure ranges from 1% to 5% (Mendes-­Moreira et al., in the genetic diversity levels of the two populations, and (iii) evaluate
2009). Generally, the postharvest selection is the only selection that the effect of stratified mass selection in quality traits (related to con-
the farmer traditionally carries out (nonformal selection) and the one sumer preferences, technological, nutritional, and organoleptic prop-
that had been applied to the historical maize populations previously to erties) that may influence maize bread quality.
their introduction in this participatory program.
As recently reviewed by Fu (2015), besides aiming at the improve-
2 | MATERIALS AND METHODS
ment of yield, adaptation, resistance to biotic and abiotic stresses, and
end-­use quality, understanding and evaluating the impacts of (modern)
2.1 | Populations’ origin and main features
plant breeding on crop genetic diversity is crucial to face the challenges
of creating better crops/varieties capable of mitigating the constraints The two historical open-­pollinated maize populations evaluated in this
of fluctuating edaphic–climatic conditions. Moreover, genetic diversity study were previously subjected to on-­farm stratified mass selection
 |
4       ALVES et al.

in the context of a participatory breeding program. This breeding due to a reduction in the breeding activities at the Lousada site, the
program has been running in Portugal since 1984 in the Sousa Valley Castro Verde population began to be selected at the Coimbra site
region, in the northern part of the country. Each maize population in (40°13′0.22″N, 8°26′47.69″W) by the breeder. At that point, some
this breeding program occupied, on average, an area of 1,000 m2 and adjustments were made to the breeding objectives but keeping the
was composed of approximately 5,000 individuals per growing season same selection methodologies (stratified mass selection). Specifically,
(given a plant density of 50,000 plants/ha). selection criteria were fine-­tuned to decrease the height of the ear in-
Amiúdo, a yellow flint early population (FAO 200), was chosen to sertion on the stalk, increase the stalk resistance, and keep increasing
integrate the PPB program in its beginning, in 1984. This population the ear size while still maintaining an orange flint kernel.
was selected due to its short life cycle and because it had already As a result of 19 years of Amiúdo selection at Lousada site, 19
adapted to the local conditions (poor soils with low pH, water stress, cycles of stratified mass selection were originated, and as a result of
and aluminum toxicity); it was also chosen because it could be used for 25 years of Amiúdo selection at Serra do Carvalho site, 25 cycles of
bread production (Vaz Patto et al., 2013). stratified mass selection were originated. In this study, the follow-
Amiúdo population was selected at two different locations: at the ing Amiúdo cycles were analyzed: the initial population from 1984,
Lousada site (41°14′7.8″N 8°18′11.1″W), where the selection was considered as cycle 0 (hereafter referred to as AMC0-1984), and the
performed by the breeder and farmer; and at the Serra do Carvalho nineteenth and the twenty-­fifth cycles of stratified mass selection,
site (41°34′12.74″N, 8°19′28.77″W), where the selection was per- obtained in 2003 at the Lousada site (hereafter referred to as AM-­
formed by the breeder. In both cases, the specific breeding objective, LC19-2003) and in 2009 at the Serra do Carvalho site (hereafter referred
set by the farmer, was to achieve a higher-­yielding population; the to as AM-­SCC25-2009), respectively.
same selection methodologies were applied at both the Lousada and As a result of 14 years of Castro Verde selection, 14 cycles of strat-
Serra do Carvalho sites. ified mass selection were originated between Lousada and Coimbra
Castro Verde, an orange flint late population (FAO 600), was intro- sites. In this study, the following Castro Verde cycles were analyzed:
duced in the PPB program in 1994 with the initial aim of achieving a the initial population from 1994, considered as cycle 0 (hereafter re-
population that could run in the category of yellow flint in a contest for ferred to as CAC0-1994), and the ninth and fourteenth cycles of strat-
the “Best Ears” of the Sousa Valley. This population was characterized ified mass selection at Coimbra obtained in 2004 (hereafter referred
by its big ears and very tall plants (>3 m in height). to as CA-­CC09-2004) and in 2009 (hereafter referred to as CA-­CC14-2009),
Until 2000, Castro Verde was selected at the Lousada site respectively.
(41°14′7.8″N 8°18′11.1″W) by the farmer. The selection criteria The summary of the specific breeding objectives for the Amiúdo
were set to obtain bigger ears by improving the traits that might en- and Castro Verde populations, as well as the timeline and selection
able the ears to win the “Best Ears” contest, namely ear length and sites where the different cycles, analyzed in this work, were devel-
kernel weight, row number, and number of kernels per ear. After 2001, oped, is given in Figure 1.

F I G U R E   1   Breeding objectives, timeline, and selection sites for the analyzed Amiúdo cycles (initial population—AMC0-1984; AM-­LC19-2003
selection cycle; and AM-­SCC25-2009 selection cycle) and Castro Verde cycles (initial population—CAC0-1994; CA-­CC09-2004 selection cycle; and CA-­
CC14-2009 selection cycle)
ALVES et al. |
      5

analysis (PCA) on the standardized agronomic data was performed

2.2 | Agronomic evaluation
using the PROC PRINCOMP procedure. The number of principal
The agronomic performance of two historical maize populations, components was determined by inspecting eigenvalues of principal
Amiúdo and Castro Verde, and their derived selection cycles was components (using the Kaiser criterion that retains components with
compared in multilocation field trials. The Amiúdo initial population eigenvalues greater than one). The first two principal components
(AMC0-1984) and selection cycles (AM-­LC19-2003 and AM-­SCC25-2009) were then projected in a biplot to display shifts occurring in the agro-
were evaluated in eight locations: Quinta da Conraria, Montemor-­ nomic traits measured on both initial populations and their selection
o-­Velho, S. Pedro do Sul, Lousada, Valada do Ribatejo, Vouzela-­1, cycles.
Vouzela-­2, and Travassos. The Castro Verde initial population
(CAC0-1994) and selection cycles (CA-­CC09-2004 and CA-­CC14-2009) were
2.4 | Molecular evaluation
evaluated in five locations: Quinta da Conraria, Montemor-­o-­Velho,
Lousada, Valada do Ribatejo, and Covas do Monte. The different Thirty random individual plants from the Amiúdo and Castro Verde
locations represent different areas where maize open-­pollinated initial populations and derived selection cycles were genotyped with
populations are traditionally produced in the country and also the 20 microsatellites (SSRs—simple sequence repeats). SSRs were cho-
different agronomic production systems normally associated with sen based on their location in the maize reference genome (1 SSR per
maize open-­pollinated populations, ranging from conventional chromosome arm) and repeat motifs (≥3 base pairs) to facilitate al-
production systems (Montemor-­o-­Velho) to organic production lele scoring (Table S2). Information about each SSR can be found at
systems (Quinta da Conraria and Valada do Ribatejo) to low-­input MaizeGDB (Lawrence et al., 2008, www.maizegdb.org).
production systems (all the other locations). Information about the DNA was isolated from adult leaves of each plant using the
sites’ characterization is given in Table S1. Initial populations and modified CTAB procedure as described in Saghai-­Maroof, Soliman,
selection cycles were evaluated, at farmers’ fields, in a randomized Jorgensen, and Allard (1984). DNA quality was accessed using a 0.8%
complete block design, with three blocks per location. Each initial SeaKem® LE Agarose gel (Cambrex Bio Science Rockland, Inc., USA)
population and derived selection cycles were overplanted by hand in stained with SYBR® Safe (Invitrogen, USA). DNA quantification was
two-­row plots 6.4 m long and with 0.75 m between rows. Each plot performed using a spectrophotometer, Nanodrop ND-­2000C (Thermo
was thinned at the seven-­leaf stage to 48 plants per plot to achieve Scientific, USA). An additional step for polysaccharide removal (Rether,
a plant density of 50,000 plants/ha. Therefore, in each environment Delmas, & Laouedj, 1993) was added when the ratio 260/230 nm
a total of 144 plants (48 plants per plot*3 blocks) were evaluated for wavelength was inferior to 1.6 to avoid the interference of these con-
each cycle. Plots were irrigated as needed and mechanically weeded taminants in SSR amplification.
and/or hand-­weeded as necessary. All the plots were harvested by The SSR loci were amplified using a nested-­PCR method (Schuelke,
hand. 2000). PCR products were separated on 6.5% polyacrylamide se-
The agronomic evaluation of each initial population and derived quencing gel (20 μl 6.5% KBPlus Gel Matrix, 150 μl APS 10%, and 15 μl
selection cycles was performed as described in Table 1. The data col- TEMED) using a LI-­COR 4300 DNA analyzer system. To account for
lected were intended to track eventual changes occurring in ear mor- any variance between PCR amplifications and electrophoresis runs,
phology, plant architecture, plant health and quality of the stalk and DNA from the B73 maize inbred line was used as a reference sam-
root system, population uniformity, and grain production. ple. Scoring of the alleles was confirmed manually by two independent
users to insure scoring accuracy. A genotypic matrix of the alleles per
individual plant, scored in base pairs, was generated and served as the
2.3 | Agronomic data analysis
basis for the molecular data analysis.
All agronomic data analysis was carried out in SAS software (SAS
Release 9.2.; SAS Institute, 2004).
2.5 | Molecular data analysis
Analysis of variance for Amiúdo cycles (initial population—AMC0-
1984; AM-­LC19-2003 selection cycle; and AM-­SCC25-2009 selection cycle) To assess the intracycle genetic diversity, the average number of al-
and for Castro Verde cycles (initial population—CAC0-1994; CA-­CC09- leles per locus (Nav), observed (HO) and expected heterozygosity (HE),
2004 selection cycle; and CA-­CC14-2009 selection cycle) was carried and inbreeding coefficient (FIS) were calculated for each initial popula-
out separately per population using the PROC MIXED procedure. In tion and selection cycles using GENEPOP software (GENEPOP v4.0;
the mixed-­model statement, environments and cycles (initial popula- Raymond & Rousset, 1995). The values of these estimates, obtained
tion and derived selection cycles) were treated as fixed effects, while in each initial population and selection cycles, were then compared to
blocks, treated as random, were nested in the environments. The in- test whether the values of Nav, HO, HE, and FIS were significantly dif-
teraction between cycles and the environment was included in the ferent among cycles with the Kruskal–Wallis test using SAS software
model. Cycle means were compared using a Tukey–Kramer multiple (SAS Release 9.2, SAS Institute Inc 2004).
comparisons test. The genotypic frequencies for each locus and for each Amiúdo and
To summarize multivariate changes occurring in both populations Castro Verde cycles were tested for conformance to Hardy–Weinberg
across the participatory breeding program, a principal component (HW) expectations using GENEPOP software (GENEPOP v4.0;
 |
6       ALVES et al.

T A B L E   1   List of agronomic traits evaluated per plot basis, codes, and respective description

Type of trait Trait Code Units/Scale Description

Ear morphology Ear weight EW Gram (g) Ear weight, adjusted to 15% of grain moisture. Measured by
averaging the weight of 4 shelled ears per plot.
Cob weight CW Gram (g) Cob weight, adjusted to 15% of grain moisture. Measured by
averaging the weight of the cobs of 4 shelled ears per plot.
Cob weight/ear CWEW Ratio (g/g) Ratio cob/ear weight indicates the proportion of cob weight in
weight the ear weight. This ratio was taken from the cob and ear
weights of 4 shelled ears per plot.
Ear moisture EM Percentage (%) Measured with a FARMPOINT moisture meter, using a mixture
sample of 4 shelled ears grain per plot.
Plant architecture Ear placement E 1–9 scale Ear placement in the plant. In this scale, a 5 indicates that the
first ear is located in the middle of the plant; and values <5
indicate that the first ear is located bellow the plant middle
point; values >5 indicate that the first ear is located above the
plant middle point. This value was measured by evaluating all
plants per plot.
Leaf angle N 1–9 scale Angle of the adaxial side of the leaf above the ear with the stalk.
In this scale, a 5 indicates a leaf angle = 45 °; values <5 indicate
a leaf angle <45°; and values > 5 indicate a leaf angle >45°.
This value was measured by evaluating all plants per plot.
Tassel branching T 1–9 scale In this scale, 1 indicates unbranched tassel (typical of inbred
lines) and 9 indicates a highly branched tassel (frequent in
populations with fasciated ears). This value was measured by
evaluating all plants per plot.
Health and quality of Root lodging R Percentage (%) Root lodging corresponds to percentage of plants leaning more
the stalk and root than 30° from vertical in each plot. This value was measured by
system evaluating all plants per plot.
Stalk lodging S Percentage (%) Stalk lodging corresponds to percentage of plants broken at or
below the primary ear node. This value was measured by
evaluating all plants per plot.
Standing plants SP No. plants/ Estimation of the number of standing plants per hectare given
hectare (no. the number of plants at harvest time in the area of each plot
plants/ha) (9.6 m2).
Population Uniformity U 1–9 scale Measure of population uniformity. In this scale, 1 indicates
uniformity minimum uniformity and 9 indicates maximum uniformity.
Values from 1 to 4 are typical of open-­pollinated populations,
and values from 5 to 9 are typical of pure lines. Measured by
evaluating all plants per plot.
Grain production Prolificacy P No. ears/plant Total number of ears per plot divided by the total number of
plants per plot.
Grain yield Y Kilogram/hectare Grain yield adjusted to 15% moisture. Formula: Grain yield = Ear
(kg/ha) weight × (Grain weight/Ear weight) × (100%–% moisture at
harvest)/(100%–15% moisture). Grain weight and ear weight
taken from 4 shelled ears.
Grain yield per plant YP Gram/plant Grain yield adjusted to 15% moisture divided by the number of
(g/plant) standing plants per hectare.

Raymond & Rousset, 1995). The probability test was based on the can be explained entirely by genetic drift and sampling error. For
Markov chain method (Guo & Thompson, 1992; Raymond & Rousset, the Amiúdo population, the temporal variation in allele frequencies
1995) using 10,000 dememorization steps, 20 batches, and 5,000 iter- was tested (i) between the Amiúdo initial population (AMC0-1984)
ations per batch. The sequential Bonferroni adjustments (Rice, 1989) and the selection cycle from the Lousada site (AM-­LC19-2003), and
were then applied to correct for the effect of multiple tests using SAS (ii) between the Amiúdo initial population (AMC0-1984) and the selec-
software (SAS Release 9.2, SAS Institute Inc 2004). tion cycle from the Serra do Carvalho site (AM-­SCC25-2009). For the
Differences in allele frequencies distributions along the breeding Castro Verde population, the temporal variation in allele frequencies
program were tested according to Waples (1989a), in which the null was tested between the initial Castro Verde population (AMC0-1984)
hypothesis states that the observed differences in allele frequency and the latter selection cycle from the Coimbra site (CA-­CC14-2009).
ALVES et al. |
      7

Afterward, the sequential Bonferroni adjustments (Rice, 1989) were

2.6.1 | Pasting behavior
applied to the level of significance to correct for the effect of mul-
tiple tests using SAS software (SAS Release 9.2, SAS Institute Inc The pasting properties of maize flour were obtained with a Rapid
2004). The effective population size, which is a parameter neces- Viscosity Analyzer RVA-­4 (Newport Scientific, Australia) at 15% solids
sary to test for temporal variation in allele frequencies, according as described in Brites et al. (2010). Peak (PV), minimum or trough (TV),
to Waples (1989a), was estimated using NeEstimator software and final viscosities (FV) were recorded in cPoise, and the breakdown
(NeEstimator v2.01, Do et al., 2014) following the temporal-­based (BD) was calculated as PV-­TV.
method under sample plan II (Waples, 1989b), as the samples an-
alyzed did not return to the breeding program. Alleles with a fre-
2.6.2 | Flour color parameters
quency lower than 0.05 were excluded, parametric chi-­squared 95%
confidence intervals for effective population size were calculated, Maize flour color was determined on 10–12 g of sample in an opaque
and the variance in allele frequencies was calculated according to recipient using a Minolta chromameter CR-­2b and CIE tristimulus
Nei and Tajima (1981). color parameters: L*—lightness; a*—red/green index; and b*—yellow/
Analysis of molecular variance (AMOVA; Excoffier, Smouse, & blue index. L* values can vary from L* = 0 (black) to L* = 100 (white);
Quattro, 1992), which is a method of estimating population differen- positive a* values mean that samples tend toward the red part of the
tiation directly from molecular data, was used to test whether the dif- color spectra; and positive b* values mean that samples tend toward
ferent cycles from Amiúdo and Castro Verde populations had suffered the yellow part of the color spectra.
genetic differentiation along the breeding program. This was done by
testing the partition of the total microsatellite diversity between and
2.6.3 | Protein, fat, and fiber content
within each pair of cycles, as well as among and within all cycles using
ARLEQUIN software (ARLEQUIN v3.0; Excoffier, Laval, & Schneider, Flour protein (PR), fat (FT), and fiber (FI) content were determined by
2005). The variance components retrieved from AMOVA were used a near-­infrared spectroscopic method with an Inframatic 8620 equip-
to calculate a series of statistics called ϕ-­statistics, which summarize ment (Perten, Sweden), with calibrations supplied by the manufac-
the degree of differentiation between population divisions and are turer. Results were expressed in percentage.
analogous to Wright’s F-­statistics (Excoffier et al., 1992). The variance
components were tested statistically by nonparametric randomization
2.6.4 | Total carotenoid content
tests using 10,000 permutations in ARLEQUIN software (ARLEQUIN
v3.0, Excoffier et al., 2005). The total carotenoid content (TCC) was spectrophotometrically meas-
To represent genetic relationships among individual plants, a fac- ured at 450 nm according to the AACCI method 14-­60.01 (AACC
torial correspondence analysis (FCA) was carried out using GENETIX International 2012). Results were expressed in μg of lutein equivalent
software (GENETIX v4.05; Belkhir, Borsa, Chikhi, Raufaste, & per gram of sample, as the main carotenoid found in maize.
Bonhomme, 2004), as this analysis provides a way of visually showing
how genetically distant the different initial populations and derived
2.6.5 | Tocopherols content
selection cycles are; it also serves as a method for observing the level
of genetic homogeneity within each cycle. α-­Tocopherol (AT), γ-­tocopherol (GT), and δ-­tocopherol (DT)
were separated from the fat portion of the maize flours by high-­
performance liquid chromatography (HPLC) and quantified using
2.6 | Quality evaluation
an Agilent 1200 model with a fluorescence detector (FLD) and
As both populations are used for human consumption, we also a Diol column (LiChropher 100, 250 × 4 mm) according to the
measured in each of the Amiúdo and Castro Verde initial popula- method ISO 9936 (2006). Tocopherols content was expressed in
tions and derived selection cycles several traits associated with ker- μg/g fat basis.
nel quality. Therefore, this study also intended to evaluate in which
way traits related to flour’s pasting behavior (flour viscosity param-
2.6.6 | Total free phenolic content
eters), nutritional value (protein, fat, and fiber content), potential
bioactive compounds (carotenoids, tocopherols, total phenolic Ethanolic extracts (EtOH:H2O 50:50, v/v) for assessing the total phe-
compounds content), and aroma-­related compounds (volatile alde- nolic content (PH) of maize flour were prepared as described in Lopez-­
hydes) have changed or were maintained along the PPB program. Martinez et al. (2009), with some modifications as described in detail
For that, a bulk of kernel from each selection cycle produced from a in Supplementary Material.
common-­garden experiment established in Coimbra in 2009, under Total free phenolic content was assessed using the Folin–Ciocalteu
controlled pollinations, was used. assay (Singleton, Orthofer, & Lamuela-­Raventos, 1999) with a Beckman
Wholemeal maize flour was obtained after milling the kernel DU-­70 spectrophotometer, with slight modifications as described in
through a Falling number 3100 mill (Perten, Sweden), using a 0.8-­mm Silva et al. (2015), and expressed in mg of gallic acid equivalents/100 g
screen. of dry weight (GAE/100 g DW).
 |
8       ALVES et al.

In relation to the Castro Verde population, on-­farm stratified mass

2.6.7 | p-­Coumaric and ferulic acid content
selection did not lead to any significant differences in the mean val-
p-­Coumaric (CU) and ferulic acid (FE) were quantified by HPLC cou- ues of the agronomic traits evaluated in this work (Table 3). For both
pled with a photodiode array detector (HPLC-­PDA) at 280 nm with a Amiúdo (Table 2) and Castro Verde (Table 3), no significant genotype
Thermo Finnigan Surveyor HPLC system according to Silva, Gomes, x environment interaction was detected for the agronomic traits
Leitão, Coelho, and Vilas Boas (2006). p-­Coumaric (CU) and ferulic evaluated.
acid contents were expressed in mg/100 g of dry weight (mg/100 g A principal component analysis based on the agronomic data was
DW). used to summarize the multivariate changes occurring in both pop-
ulations across the participatory breeding program. The first two
principal components for both the Amiúdo and Castro Verde cycles
2.6.8 | Volatile aldehydes content
retained 94.49% of the total variance, with the first component al-
The volatile fraction of maize flour was analyzed by solid-­phase micro- ready retaining 84.37% of observed variance (Figure 2). In the PCA
extraction–gas chromatography–mass spectrometry (SPME-­GC-­MS). biplot (Figure 2), the first axis clearly separated the Amiúdo from the
A 2-­cm 50/30-­μm DVB/Carboxen/PDMS fiber (SUPELCO) was used Castro Verde populations. Moreover, for Amiúdo the first axis sep-
for solid-­phase microextraction. Volatile compounds were analyzed arated the initial population (AMC0-1984) from the two selection cy-
with a GCMS-­QP2010 Plus Shimadzu equipment and separated in a cles (AM-­LC19-2003 and AM-­SCC25-2009) in the direction of an increase
Varian Factor Four column (30 m × 0.25 mm × 0.25 μm). Volatile alde- in all the traits analyzed except for plant prolificacy (P) and the angle
hydes content (AL) was taken as the sum of the peak area of the main of the leaf insertion in the stalk (N) that decreased in this direction.
aldehydes identified (hexanal, heptenal, 2-­heptanal (Z), 2-­octenal (E), The second axis separated the two selection cycles, AM-­LC19-2003 and
nonanal, 2-­nonenal (E), and decanal). Details on the quantification of AM-­SCC25-2009, in the direction of an increase in the number of plants
volatile aldehydes content can be found in Supplementary Material. standing (SP), with the selection cycle from the Serra do Carvalho site
having a higher number of plants standing. As for Castro Verde, and
as expected by the results obtained previously for the analysis of vari-
2.7 | Quality data analysis
ance (Table 3), no clear progression was observed along the selection
To summarize the eventual multivariate changes on the evaluated process when comparing the position on the biplot of the initial pop-
quality traits occurring in both populations across the participatory ulation CAC0-1994, the cycle from 2004 (CA-­CC09-2004), and the cycle
breeding program, a principal component analysis (PCA) was per- from 2009 (CA-­CC14-2009) (Figure 2).
formed using the PROC PRINCOMP procedure after standardization
of the quality traits, similar to what has been already described for the
3.2 | Molecular diversity evolution
agronomic data analysis.

3.2.1 | Intrapopulation diversity
3 |  RESULTS The molecular diversity analysis allowed tracing the overall genetic di-
versity evolution in the two open-­pollinated populations under study.
In this work, the agronomical, molecular, and quality evolution of In terms of quantitative differences in the alleles detected for the
two historical open-­pollinated maize populations, Amiúdo and Castro Amiúdo population, 73.26% of all alleles were maintained throughout
Verde, across a participatory plant breeding program was accessed. the cycles: Of the 86 alleles detected, 63 were common to all the cy-
cles (Table S3). Only six to eight alleles (7%–9.3%), out of the 74 iden-
tified in the initial population (AMC0-1984), were not detected in the
3.1 | Agronomic evolution
Serra do Carvalho (AM-­SCC25-2009) and in the Lousada (AM-­LC19-2003)
In relation to the Amiúdo population agronomic performance, on-­ selection cycles, respectively (Table S2). Likewise, in terms of quan-
farm stratified mass selection led, in both selection sites—Lousada titative differences in the alleles detected for Castro Verde popula-
and Serra do Carvalho—to a significant increase in ear (EW) and cob tion, the majority of the alleles (65.91%) were maintained throughout
weight (CW) and cob/ear weight ratio (CWEW) (0.9%–1.2% for EW, the cycles: Of 88 alleles detected, 58 were common to all the cycles
2.1%–3% for CW, and 1%–1.6% gain per cycle for CWEW, respec- (Table S3). Only 10 alleles (11.4%), out of the 74 detected in the ini-
tively) as well as to a significant gain in grain yield per plant (0.9% gain tial population, were not detected in the CA-­CC14-2009 selection cycle
per cycle) and in grain yield overall (0.8% gain per cycle) (Table 2). The (Table S2).
Amiúdo selection cycle from the Lousada site also had a significant As for the allelic frequencies, for both Amiúdo and Castro Verde
increase in the levels of ear moisture (0.5% gain per cycle) when com- populations a considerable proportion of the alleles detected were
pared with the initial population (Table 2). The selection performed at present in low frequencies (0.1 or less): Amiúdo cycles with 39.19% at
the Serra do Carvalho site gave rise to an Amiúdo population with a the initial population (AMC0-1984), 41.89% at the selection cycle from
decreased percentage of stalk lodging (−1.4% gain per cycle), and to the Lousada site (AM-­LC19-2003), and 48.10% at the selection cycle from
an increase in tassel branching (0.4% gain per cycle) (Table 2). the Serra do Carvalho site (Fig. S1A); and Castro Verde cycles with
ALVES et al. |
      9

T A B L E   2   Analysis of variance, comparison of mean values, and percentage of gain per selection cycle for the agronomic traits among
Amiúdo initial population (AMC0-1984) and selection cycles from Lousada (AM-­LC19-2003) and Serra do Carvalho (AM-­SCC25-2009)

Comparison of initial population/selection % gain/ % gain/

Analysis of variancea cycle meansb cycle cycle

AM-­ AM-­
Trait Cycle Env Cycle*Env AMC0-1984 AM-­LC19-2003 AM-­SCC25-2009 LC19-2003 SCC25-2009

Ear weight (EW), in g *** ns ns 124.35 b 146.07 a 162.56 a 0.9 1.2

Cob weight (CW), in g *** * ns 20.29 b 31.85 a 30.83 a 3.0 2.1
Cob weight/ear weight (CWEW), *** *** ns 0.16 b 0.21 a 0.20 a 1.6 1.0
in g/g
Ear moisture (EM), in % * *** ns 18.84 b 20.59 a 20.13 ab 0.5 —
Ear placement (E), in 1–9 scalec ns ns ns 5.54 a 5.29 a 5.38 a — —
d
Leaf angle (N), in 1–9 scale ns ** ns 5.42 a 5.25 a 5.29 a — —
Tassel branching (T), in 1–9 scalee * *** ns 6.21 a 6.44 ab 6.79 a — 0.4
Root lodging (R), in % ns * ns 5.48 a 7.29 a 6.32 a — —
Stalk lodging (S), in % * ** ns 9.81 a 9.53 a 6.30 b — −1.4
Standing plants (SP), in no. ns ** ns 49236 a 50062 a 49996 a — —
plants/ha
Uniformity (U), in 1–9 scalef ns ns ns 3.42 a 3.58 a 3.38 a — —
Prolificacy (P), in no. ears/plant ns ns ns 1.07 a 1.10 a 1.05 a — —
Grain yield (Y), in kg/ha ** ns ns 4568.84 b 5322.79 a 5577.93 a 0.8 0.8
Grain yield per plant (YP), in ** ns ns 93.00 b 107.88 a 112.57 a 0.9 0.9
g/plant
a
Significance for analysis of variance among cycles (initial population plus selection cycles) and among environments (Env) and interaction between cycles
and environments (Cycle*Env): ns—nonsignificant; *—significant at p < .05; **—significant at p < .01; ***—significant at p < .001.
b
Tukey–Kramer multiple comparisons test—mean values in each row followed by the same letter are not significantly different at p < .05.
c
Ear placement (E), in 1–9 scale: 5 indicates that the first ear is located in the middle of the plant; values <5 indicate that the first ear is located bellow the
plant middle point; and values >5 indicate that the first ear is located above the plant middle point.
d
Leaf angle (N), in 1–9 scale: 5 indicates a leaf angle = 45 °; values <5 indicate a leaf angle <45 °; and values >5 indicate a leaf angle >45 °.
e
Tassel branching (T), in 1–9 scale: 1 indicates unbranched tassel and 9 indicates a highly branched tassel.
f
Uniformity (U), in 1–9 scale: 1 indicates minimum uniformity and 9 indicates maximum uniformity.

47.30% at initial population (CAC0-1994), 48.61% at the CA-­CC09-2004 (Ne = 243.7) (Table S5). For Castro Verde, the estimated effective pop-
selection cycle, and 50% at the CA-­CC14-2009 selection cycle (Fig. S1B). ulation size was Ne = 161.7 (Table S5). After a Bonferroni multiple-­test
When testing for significant differences among cycles within each correction, no significant temporal variation of allele frequencies was
population in the average number of alleles detected, observed and detected for both populations and selection sites (Amiúdo: Table S6;
expected heterozygosity, and inbreeding coefficients, no significant Castro Verde: Table S7).
differences were observed among the cycles for both the Amiúdo and
Castro Verde populations (Table 4).
3.2.2 | Differentiation among cycles
The global Hardy–Weinberg equilibrium test detected a significant
departure from Hardy–Weinberg equilibrium in the Amiúdo cycle, The genetic differentiation among cycles within each population was
AM-­SCC25-2009, and in the Castro Verde cycle, CA-­CC14-2009, both due tested following the framework of AMOVA. The AMOVA results
to heterozygote deficiency (FIS = 0.042, p-­value <.01; and FIS = 0.082, showed that for the Amiúdo population, the percentage of variance
p-­value <.05, respectively) (Table 4). When testing for the departure that could be attributed to differences among all cycles represented
from Hardy–Weinberg equilibrium by individual locus in both the 2.86% of the total molecular variation (Table 5). The pairwise com-
Amiúdo and Castro Verde populations, the majority of the loci had parisons between Amiúdo cycles showed that stratified mass selec-
their genotypic frequencies in accordance with Hardy–Weinberg ex- tion led overall to a significant but small genetic differentiation (given
pectations (Table S4). the significant ϕST values; Table 5). For the Castro Verde population,
With the objective of testing for temporal changes in the allele AMOVA showed that the variation among all cycles represented only
frequencies distribution, the effective population size (Ne) was esti- 1.72% of the total molecular variation (Table 5). In this case, strati-
mated by a temporal-­based method under sample plan II. For Amiúdo, fied mass selection did not generate a significant genetic differentia-
the estimated effective population size for the Lousada site was tion between CAC0-1994 and CA-­CC09-2004 (ϕST = 0.003, p-­value >.05)
Ne = 119.6, while for the Serra do Carvalho site the Ne value was bigger (Table 5).
10      | ALVES et al.

T A B L E   3   Analysis of variance and

Comparison of initial population/selection
comparison of mean values for the
Analysis of variancea cycle meansb
agronomic traits among Castro Verde initial
Trait Cycle Env Cycle*Env CAC0-1994 CA-­CC09-2004 CA-­CC14-2009 population (CAC0-1994) and selection cycles
(CA-­CC09-2004 and CA-­CC14-2009)
Ear weight (EW), in g ns ns ns 240.12 a 256.46 a 247.22 a
Cob weight (CW), in g ns ** ns 57.93 a 65.79 a 58.12 a
Cob weight/ear weight ns *** ns 0.24 a 0.26 a 0.23 a
(CWEW), in g/g
Ear moisture (EM), in % ns ** ns 24.20 a 24.81 a 24.20 a
Ear placement (E), in ns * ns 6.00 a 6.40 a 6.03 a
1–9 scalec
Leaf angle (N), in 1–9 ns ns ns 5.13 a 5.15 a 4.87 a
scaled
Tassel branching (T), in ns ns ns 7.07 a 7.14 a 6.97 a
1–9 scalee
Root lodging (R), in % ns ** ns 31.99 a 31.50 a 22.53 a
Stalk lodging (S), in % ns *** ns 25.20 a 25.22 a 27.93 a
Standing plants (SP), in ns ** ns 48,924 a 47,100 a 48,403 a
no. plants/ha
Uniformity (U), in 1–9 ns ** ns 3.77 a 3.80 a 3.63 a
scalef
Prolificacy (P), in no. ns ns ns 0.98 a 1.00 a 0.90 a
ears/plant
Grain yield (Y), in kg/ha ns * ns 6,862.71 a 6,851.03 a 6,840.93 a
Grain yield per plant ns ns ns 146.33 a 147.15 a 144.52 a
(YP), in g/plant
a
Significance for analysis of variance among cycles (initial population plus selection cycles) and among
environments (Env) and interaction between cycles and environments (Cycle*Env): ns—nonsignificant;
*—significant at p < .05; **—significant at p < .01; ***—significant at p < .001.
b
Tukey–Kramer multiple comparisons test—mean values in each row followed by the same letter are
not significantly different at p < .05.
c
Ear placement (E), in 1–9 scale: 5 indicates that the first ear is located in the middle of the plant; values
< 5 indicate that the first ear is located bellow the plant middle point; and values > 5 indicate that the
first ear is located above the plant middle point.
d
Leaf angle (N), in 1–9 scale: 5 indicates a leaf angle = 45 °; values < 5 indicate a leaf angle <45 °; and
values > 5 indicate a leaf angle > 45 °.
e
Tassel branching (T), in 1–9 scale: 1 indicates unbranched tassel and 9 indicates a highly branched
tassel.
f
Uniformity (U), in 1–9 scale: 1 indicates minimum uniformity and 9 indicates maximum uniformity.

variance, separated the initial cycle (CAC0-1994) from the 2004 selec-
3.2.3 | Genetic relationships among individuals
tion cycle (CA-­CC09-2004; Figure 4).
The factorial correspondence analysis depicts graphically the genetic
proximity/differentiation within and among initial populations and
3.3 | Quality evolution
selection cycles. From the factorial correspondence analysis of the
Amiúdo population, the first axis, which accounted for 66.16% of the In relation to Amiúdo quality evaluation, the breeding activities led,
observed genotypic variance, separated the initial population (AMC0- in the material developed both at Lousada (AM-­LC19-2003 cycle) and
1984) from its selection cycles. The second axis, which accounted for at Serra do Carvalho (AM-­SCC25-2009 cycle), to a slight increase in the
33.84% of the observed genotypic variance, separated the selection total carotenoid content (TCC) and in the color red/green index (a*),
cycle from the Lousada site (AM-­LC19-2003) from the selection cycle accompanied by a decrease in the levels of γ-­tocopherol (GT), protein
from the Serra do Carvalho site (AM-­SCC25-2009; Figure 3). From the (PR), fiber (FI), total volatile aldehydes (AL), total free phenolic (PH)
factorial correspondence analysis of Castro Verde, the first axis, which compounds, p-­coumaric acid (CU), and ferulic acid (FE) (Table S8).
accounted for 63.85% of the observed genotypic variance, separated In the case of Castro Verde quality evaluation, although the results
the most recent selection cycle (CA-­CC14-2009) from the other two. The showed first a reduction of the flour’s yellowness (taken as color pa-
second axis, which accounted for 36.15% of the observed genotypic rameter b* values) from CAC0-1994 to CA-­CC09-2004 and afterward from
ALVES et al. |
      11

F I G U R E   2   Biplot of principal
component analysis (PCA) based on 14
agronomic traits measured in the Amiúdo
cycles (initial population—AMC0-1984; AM-­
LC19-2003 selection cycle; and AM-­SCC25-2009
selection cycle) and Castro Verde cycles
(initial population—CAC0-1994; CA-­CC09-2004
selection cycle; and CA-­CC14-2009 selection
cycle)

T A B L E   4   Genetic variability estimates

p-­Value
for Amiúdo initial population (AMC0-1984)
Population/Selection cycle N Nav Npr HO HE FIS HWE
and Castro Verde initial population
(CAC0-1994) and derived selection cycles AMC0-1984 30 3.70 3 0.537 0.532 −0.009 ns
AM-­LC19-2003 30 3.70 1 0.523 0.531 0.014 ns
AM-­SCC25-2009 30 3.95 4 0.503 0.526 0.042 **
p-­valuea (KW) .961 .584 .725 .520
CAC0-1994 30 3.70 4 0.482 0.482 0.000 ns
CA-­CC09-2004 30 3.60 2 0.456 0.482 0.054 ns
CA-­CC14-2009 30 3.80 6 0.457 0.498 0.082 *
p-­valuea (KW) .911 .790 .930 .825

N, number of individuals; Nav, average number of alleles; Npr, number of private alleles; HO, observed
heterozygosity; HE, gene diversity or expected heterozygosity; FIS, inbreeding coefficient; p-­value
HWE, The probability global test for Hardy–Weinberg equilibrium (HWE) for each cycle was based on
Markov chain method; ns, nonsignificant.
a
p-­Value of Kruskal–Wallis test among cycles (initial populations and derived selection cycles).
*Significant at p < .05; **Significant at p < .01; ***Significant at p < .001.

CA-­CC09-2004 to CA-­CC14-2009 cycle, the b* value stopped decreasing. biplot revealed an increase in the levels of α-­ and δ-­tocopherol (AT
Moreover, it was observed an increase in the levels of (α-­, δ-­, and γ-­) and DT) and fat (FT) when comparing the Amiúdo initial population
tocopherols (AT, DT, GT), and p-­coumaric acid (CU), as well as a de- (AMC0-1984) with the Amiúdo cycle from the Lousada selection site
crease in the levels of fiber (FI), protein (PR), and total free phenolic (AM-­LC19-2003). While comparing the Amiúdo initial population (AMC0-
(PH) compounds along the selection cycles. Nevertheless, for Castro 1984) with the Amiúdo cycle from the Serra do Carvalho selection site
Verde the majority of the quality traits (10 of 18) variation was erratic (AM-­LC25-2009), an opposite trend was depicted with a decrease in the
along selection cycles. levels of α-­ and δ-­tocopherol (AT and DT), and fat (FT), accompanied
As for the principal component analysis based on the quality data by a decrease in levels of p-­coumaric acid (CU).
in both the Amiúdo and Castro Verde populations, the first two com-
ponents retained 73.20% of the total observed variance, with the first
component explaining 50.99% of the observed variance (Figure 5). 4 | DISCUSSION
The traits that primarily influenced the first component were α-­ and
δ-­tocopherol (AT and DT), fat (FT), peak and trough viscosities (PV and Amiúdo and Castro Verde are two historical open-­pollinated maize
TV), and protein content (PR). The trait that primarily influenced the populations that have been subjected to on-­farm stratified mass se-
second component was the p-­coumaric acid (CU) content. The PCA lection, in the context of a long-­term participatory breeding program.
 |
12       ALVES et al.

T A B L E   5   Analysis of molecular variance

% Total variance
(AMOVA) results for the partitioning of
Comparison Among cycles Within cycles ϕ-­statisticsa p (ϕ)b SSR variation among and within Amiúdo
cycles (AMC0-1984, AM-­LC19-2003, and
AMC0-1984 vs. AM-­LC19-2003 4.33 95.67 0.043 ***
AM-­SCC25-2009) and Castro Verde cycles
AMC0-1984 vs. AM-­SCC25-2009 2.98 97.02 0.030 *** (CAC0-1994, CA-­CC09-2004, and CA-­CC14-2009)
AM-­LC19-2003 vs. AM-­SCC25-2009 1.22 98.78 0.012 *
All Amiúdo cycles 2.86 97.14 0.029 ***
CAC0-1994 vs. CA-­CC09-2004 0.34 99.66 0.003 ns
CAC0-1994 vs. CA-­CC14-2009 2.40 97.60 0.024 ***
CA-­CC09-2004 vs. CA-­CC14-2009 2.36 97.64 0.024 ***
All Castro Verde cycles 1.72 98.28 0.017 ***
a
ϕ-­statistics: corresponding to an analogous to the fixation index (FST) which measures the degree of
genetic differentiation among populations/selection cycles (ϕST).
b
p(ϕ): The level of significance of the ϕ-­statistics was tested by nonparametric randomization tests
using 10,000 permutations.
ns, nonsignificant; *Significant at p < .05; **Significant at p < .01; ***Significant at p < .001.

F I G U R E   3   Factorial correspondence
analysis (FCA) of 90 maize plants belonging
to the Amiúdo cycles (initial population—
AMC0-1984; AM-­LC19-2003 selection cycle;
and AM-­SCC25-2009 selection cycle). Each
individual genotype is indicated by a small
symbol, while the cycle’s mean value is
represented by larger ones

The results presented here revealed that this participatory program is

4.1 | Phenotypic effects of stratified mass selection
improving or maintaining yield and quality parameters, while preserv-
ing genetic diversity of maize populations. Additionally, this program is The results obtained from multilocation field trials, established to eval-
empowering farmers as they keep the decision power and learn some uate the effects of stratified mass selection in these two maize popula-
basic population improvement methodologies, and at the same time tions, showed that this methodology was able to improve the Amiúdo
represents an alternative strategy for endangered genetic resources’ population, according to the established selection criteria in two dif-
on-­farm conservation. ferent selection sites (Lousada and Serra do Carvalho). Nevertheless,
ALVES et al. |
      13

F I G U R E   4   Factorial correspondence
analysis (FCA) of 90 maize plants
belonging to the Castro Verde cycles
(initial population—CAC0-1994; CA-­CC09-2004
selection cycle; and CA-­CC14-2009 selection
cycle). Each individual genotype is indicated
by a small symbol, while the cycle’s mean
value is represented by larger ones

according to the data collected, the same methodology failed to lead Nevertheless, an analysis of most of the other breeding criteria—
to an agronomic improvement of the Castro Verde population. achieve bigger ears, decrease the height of the ear insertion in the
The Amiúdo population, integrated on the PPB program since its plants, and increase stalk resistance—showed that no significant im-
beginning, was selected by two different people, in two different se- provements were obtained for the Castro Verde population using this
lection sites, but with similar edaphic–climatic conditions. For both se- methodology.
lection sites, achieving a higher-­yielding population was the breeding
objective established by the farmer. Indeed, Amiúdo population had
4.2 | Implications for a quality-­oriented
a yield increase through mass selection (0.8% gain per cycle) accom-
breeding program
panied by heavier cobs and ears. This gain was however inferior to
the experimental values obtained across long-­term maize recurrent se- An important aspect of both the Amiúdo and Castro Verde popula-
lection methods for population improvement, as reviewed by Betrán, tions is the fact that their flours can be used for food. In fact, a recent
Bänziger, and Menz (2004). According to Betrán et al. (2004), when sensory hedonic analysis of maize bread, including bread obtained
grain yield is the primary selection criterion, mass selection showed from these populations, showed that both populations were able to
on average a 1.8% gain per cycle, being this value often smaller than produce bread with preferential characteristics (Carbas et al., 2016).
the average values obtained with family-­based recurrent selection, With the objective of integrating these two populations in a quality-­
such as selfed—S1 or S2—family selection (with 7% and 5% gain per oriented breeding program in due course, several traits related to
cycle, respectively). One of the reasons for the slower yield progress consumer preferences and technological, nutritional, and organolep-
observed in Amiúdo population in comparison with these reviewed tic properties (quality traits) were measured. It was observed that the
values, besides its particular genetic background, may be a reflection majority of those traits progressed erratically along the breeding pro-
of the lower selection intensity applied under the present participatory gram for the Castro Verde population. One exception was the total
program (1%–5%). carotenoid content, which can be selected efficiently by choosing the
As for Castro Verde population, the phenotypic data showed that more yellow/orange ears as the b* parameter (yellowness) is highly
stratified mass selection was able to partially induce phenotypic dif- correlated with total carotenoid content (Kljaka, Grbešaa, & Karolyib,
ferences that follow the direction of the breeding objectives (main- 2014). In general for quality traits, as the ones considered in this work,
tenance of orange grain color set as breeding criterion after 2001). a direct visual selection, like the one performed for the agronomic
 |
14       ALVES et al.

F I G U R E   5   Biplot of principal
component analysis (PCA) based on 18
quality traits in the Amiúdo cycles (initial
population—AMC0-1984; AM-­LC19-2003
selection cycle; and AM-­SCC25-2009
selection cycle) and Castro Verde cycles
(initial population—CAC0-1994; CA-­CC09-2004
selection cycle; and CA-­CC14-2009 selection
cycle)

traits, is not possible, and other complementary breeding methodolo- the lack of agronomic progress in ear architecture-­related traits in
gies are needed to encourage their effective improvement by farmers. the Castro Verde population should not be due to the trait under
selection per se. Moreover, as the analysis of variance did not de-
tect a significant genotype-­by-­environment interaction, the lack of
4.3 | Breeding program weaknesses and
Castro Verde progress should not be a consequence of this interac-
strengths analysis
tion. Instead, it could be most likely related to two particular aspects
When grain yield was the primary breeding objective, on-­farm strati- of the Castro Verde population: First, as the selection criterion until
fied mass selection, as described in this work, was effective in im- the year 2000 was set to get bigger ears, one hypothesis is that
proving population yield although at a slower rate than what can be because this population had already ears of a significant size before
obtained through other more complex family-­based recurrent selec- entering the breeding program, the farmer was not fully engaged
tion methods. With more diverse breeding objectives, as in the case with the breeding activities. Second, after 2001, this population
of Castro Verde population, stratified mass selection was not always started to be selected at Coimbra site by the breeder. Therefore,
effective in achieving the same progress. another hypothesis for the lack of observable agronomic progress is
An extensive compilation of several cases of yield improvement that the population did not have adequate isolation, as other pop-
achieved through mass selection in maize can be found at Hallauer ulations were also being grown at the same site; and the number
et al. (2010, table 7.8, therein). A few examples that show the poten- of individual plants screened may have been too small to select/
tial of stratified mass selection specifically in the context of a partic- capture the best genotypes. Indeed, Castro Verde initial population,
ipatory maize breeding program were described in Mendes-­Moreira which resulted from years of farmers traditional selection based
et al. (2008, 2009) and Smith et al. (2001). In the first two works, mainly on ear traits evaluated after harvest, had already a high
two other maize populations from the same Portuguese breeding grain yield for an open-­pollinated maize population (6,862.71 kg/
program as in the present study had their agronomic performance ha). Probably due to this, a yield increase was not the main objec-
improved in line with the farmers’ breeding objectives (Mendes-­ tive of the farmer involved on Castro Verde selection. This however
Moreira et al., 2008, 2009). Also Smith et al. (2001) showed that was not the case for the farmer involved on Amiúdo selection that
tree cycles of stratified mass selection applied to five different was aiming to improve the population initial yield (4,568.84 kg/ha).
Mexican maize populations were sufficient to obtain an increase Nevertheless, both original maize populations showed on average
in yield. Several factors have been identified as having an impact higher yields than the only data publicly available on nonimproved
on mass selection effectiveness or ineffectiveness (Hallauer et al., historical Portuguese maize populations with high quality potential
2010). Among them, one can highlight the trait under selection, an for maize bread broa production (Vaz Patto et al., 2007). Grain yield
adequate isolation, the sample size utilized, genotype x environment of these traditional populations was evaluated in a common-­garden
interaction, and the precision of the experimental techniques used field experiment, and it varied from 755 to 3,757 kg/ha, with an
(environmental control, parental control). In the present work, it was average of 1,982 kg/ha (Vaz Patto et al., 2007).
shown that the selection methodology was able to alter traits re- In the maize populations analyzed in the present study, not only
lated to ear architecture in the Amiúdo population, and therefore, natural selection but also human selection is affecting yield. In a
ALVES et al. |
      15

review by Murphy, Carter, and Jones (2013), several examples of showed that the overall genetic diversity was maintained in both pop-
the effectiveness of evolutionary breeding (accounting only for nat- ulations. In particular, even in the Amiúdo population where pheno-
ural selection) in improving the agronomic fitness of self-­pollinated typic modifications on ear morphology and yield gain were detected,
cereal crops have been examined. With this breeding approach, no significant changes were identified on the overall genetic diversity
improvement resulted from natural selection favoring high-­yielding levels, measured by the average number of alleles detected, observed
genotypes as an outcome of the relationship between the yield ca- and expected heterozygosity, and inbreeding coefficients. Also, no
pacity of an individual plant and its fitness components (Murphy significant temporal variation of allele frequencies was detected in
et al., 2013). This yield increase is highly dependent on the selec- any of populations under study, indicating that the observed differ-
tive environmental pressure and may affect maturity, plant height, ences in allele frequency are more likely a result of genetic drift and/
and relationships among agronomic important traits unfavorably or sampling error (Waples, 1989a). As opposed to the results obtained
(Phillips & Wolfe, 2005). A comparison between the yield progress by Labate et al. (1999) and Solomon et al. (2010), in which the authors
attained under the studied participatory breeding program and the detected a loss of genetic diversity in maize population subjected to
yield progress that might be attained with an evolutionary breeding few as 11 and 12 cycles of reciprocal recurrent selection, no significant
approach could have generated relevant information on the effec- differences in genetic diversity levels were identified in the current
tiveness of the human (artificial) selection versus natural selection. study. According to Hoban et al. (2014), changes in genetic diversity
Unfortunately, no references were found in the literature on the ef- levels are most likely identified only when the effective population
fect of evolutionary breeding in maize populations to allow a direct size is smaller than 100 individuals. In the present work, both popula-
comparison with the present study. However, by performing the se- tions had an effective population size bigger than 100, by contrast to
lection of Amiúdo and Castro Verde populations within the target the smaller effective population sizes estimated for the maize popula-
environment (at the farmers’ fields), on-­farm participatory breeding tions in Labate et al. (1999) and Solomon et al. (2010). In addition, the
guarantees local adaptation and it may also counteract undesirable results presented here concur with the results previously described
changes caused by natural selection in traits of agronomic impor- for the Portuguese Pigarro maize population (Vaz Patto et al., 2008)
tance. Moreover, by respecting farmers’ breeding objectives, an in- where stratified mass selection demonstrated to be an effective way
crease in the ratio of improved populations adopted by the farmer to conserve diversity on-­farm, and at the same time allowed relevant
can be obtained. phenotypic improvements to be achieved.
Although one can argue that differences in response to selection
in a similar genetic background may be due to different intensity or
4.5 | Final remarks
accuracy of selection, the acceptance and the enthusiasm of the farm-
ers to join the program are the best guaranty of success. Farmers need In conclusion, on-­farm stratified mass selection in the context of a
to be fully engaged on the selection decision process (breeding ob- participatory plant breeding program was shown to improve the ag-
jectives) but be open to accept breeder recommendations (preharvest ronomic performance of the Amiúdo population selected in two dif-
parental control + postharvest selection). ferent selection sites. Moreover, for both the Amiúdo and Castro
One open question in the present study is: How able is the farmer Verde populations, the breeding activities retained the populations’
to perform pre-­harvest trait selection? In the present work, the pre- genetic diversity. The unpredictability of the evolution of quality pa-
harvest selection was not exclusive but mainly performed by the rameters along this breeding program also brings to light the need to
breeder, and therefore, the farmer’s ability could not be clearly eval- develop efficient selection tools to maintain or improve these traits.
uated. Nevertheless, theoretically the preharvest selection methodol- Molecular markers associated with those traits and/or high through-
ogies proposed in the Portuguese participatory breeding program are put spectroscopy-­based phenotypic screening methodologies are
very straightforward and are beforehand demonstrated by the breeder among the tools that may aid in the improvement of characteristics
at the farmer’s field. Therefore, these methodologies should be easily that cannot be easily (visually) selected by farmers. The implemen-
implemented by any farmer engaged in the breeding process. Indeed, tation of such breeding tools into participatory selection brings up
it has been already demonstrated by Mendes-­Moreira et al. (2008) another issue: To make these tools easily available, a platform of par-
that such preharvest methodologies were successfully implemented ticipatory research connecting enthusiastic, open-­minded farmers,
by farmers in another maize population from the same participatory breeders, and scientists must be built to make its application a reality.
breeding program. The farmer’s motivation and time availability/field
dimensions (the bigger the field, the larger amount of time needed for
AC KNOW L ED G EM ENTS
stratified preharvest selection) seem to be the two main limitations for
the successful implementation of this preharvest selection. The research leading to these results has received funding from the
European Union’s 7th Framework Programme (FP7/2007-­2013) and
Horizon 2020 Framework Programme under grant agreement No.
4.4 | Genotypic effects of stratified mass selection
245058 and No. 633571, respectively, and from the Fundação para
The effect of stratified mass selection in the genetic diversity levels a Ciência e a Tecnologia through R&D unit, UID/Multi/04551/2013
of the two populations was also evaluated using SSRs. This analysis (GreenIT). Mara Lisa Alves and Maria Carlota Vaz Patto were supported
 |
16       ALVES et al.

by Fundação para a Ciência e a Tecnologia (SFRH/BD/77723/2011; Excoffier, L., Smouse, P. E., & Quattro, J. M. (1992). Analysis of molecu-
FCT Investigator 2015 contract IF/01337/2014, respectively). The lar variance inferred from metric distances among DNA haplotypes:
Application to human mitochondrial DNA restriction sites. Genetics,
authors would like to acknowledge Doctor Silas Pego and the farm-
131, 479–491.
ers that were involved in the development of the maize populations, Fu, Y. B. (2015). Understanding crop genetic diversity under modern plant
as well as the Portuguese Bank of Plant Germplasm for providing the breeding. Theoretical and Applied Genetics, 128, 2131–2142.
seeds from the initial landraces and selection cycles used on this work. Gardner, C. O. (1961). An evaluation of the effects of mass selection and
seed irradiation with thermal neutrons on yield of corn. Crop Science,
The authors would also like to acknowledge all the anonymous re-
1, 241–245.
viewers that contributed to the discussion and improvement of this Guo, S. W., & Thompson, E. A. (1992). Performing the exact test of Hardy-­
work. Weinberg proportion for multiple alleles. Biometrics, 48, 361–372.
Hallauer, A. R. (1999). Conversion of tropical germplasm for temperate area
use. Illinois Corn Breeders School, 35, 20–36.
DATA ARCHI VI NG S TATE M E N T Hallauer, A. R., Carena, M. J., & Miranda Filho, J. D. (2010). Quantitative
genetics in maize breeding (Vol. 6). New York, NY: Springer.
Data available from the Dryad Digital Repository: https://doi. Hellin, J., Bellon, M. R., & Hearne, S. J. (2014). Maize landraces and adap-
org/10.5061/dryad.nb320. tation to climate change in Mexico. Journal of Crop Improvement, 28,
484–501.
Hoban, S., Arntzen, J. A., Bruford, M. W., Godoy, J. A., Rus Hoelzel, A.,
O RCI D Segelbacher, G., … Bertorelle, G. (2014). Comparative evaluation of
potential indicators and temporal sampling protocols for monitoring
Mara Lisa Alves  http://orcid.org/0000-0003-3515-2082 genetic erosion. Evolutionary Applications, 7, 984–998.
ISO 9936 (2006). Animal and vegetable fats and oils – Determination
of tocopherol and tocotrienol contents by high-performance liquid
chromatography.
REFERENCES
Jarvis, D. I., Hodgkin, T., Sthapit, B. R., Fadda, C., & Lopez-Noriega, I. (2011).
AACC International (2012). Approved methods of analysis Method 14- An heuristic framework for identifying multiple ways of supporting the
60.01, 11th ed. Total carotenoid content of cereal grains and flours. St. conservation and use of traditional crop varieties within the agricultural
Paul, MN, USA: AACC International. production system. Critical Reviews in Plant Sciences, 30, 125–176.
Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N., & Bonhomme, F. (2004). Khoury, C. K., Bjorkman, A. D., Dempewolf, H., Ramirez-Villegas, J.,
GENETIX 4.05, logiciel sous Windows TM pour la génétique des pop- Guarino, L., Jarvis, A., … Struik, P. C. (2014). Increasing homogeneity in
ulations. Laboratoire Génome, Populations, Interactions, CNRS UMR global food supplies and the implications for food security. Proceedings
5000, Université de Montpellier II, Montpellier (France). of the National Academy of Sciences of the United States of America, 111,
Betrán, J., Bänziger, M., & Menz, M. (2004). Corn breeding. In C. Wayne 4001–4006.
Smith, J. Betrán, & E. C. A. Runge (Eds.), Corn: Origin, history, technology, Klensporf, D., & Jelén, H. H. (2005). Analysis of volatile aldehydes in oat
and production (pp. 305–398). New York, NY: Wiley. flakes by SPME-GC/MS. Polish Journal of Food and Nutrition Sciences,
Brites, C., Trigo, M. J., Santos, C., Collar, C., & Rosell, C. M. (2010). Maize-­ 14, 389–395.
based gluten-­free bread: Influence of processing parameters on sen- Kljaka, K., Grbešaa, D., & Karolyib, D. (2014). Reflectance colorimetry as
sory and instrumental quality. Food Bioprocess Technology, 3, 707–715. a simple method for estimating carotenoid content in maize grain.
Carbas, B., Vaz Patto, M. C., Bronze, M. R., Bento da Silva, A., Trigo, M. Journal of Cereal Science, 59, 109–111.
J., & Brites, C. (2016). Maize flour parameters that are related to the Labate, J. A., Lamkey, K. R., Lee, M., & Woodman, W. L. (1999). Temporal
consumer perceived quality of ‘broa’ specialty bread. Food Science and changes in allele frequencies in two reciprocally selected maize popu-
Technology, 36, 259–267. lations. Theoretical and Applied Genetics, 99, 1166–1178.
Ceccarelli, S. (2015). Efficiency of plant breeding. Crop Science, 55, 87–97. Lançon, J., Pichaut, J. P., Djaboutou, M., Lewicki-Dhainaut, S., Viot, C., &
Ceccarelli, S., Al-Yassin, A., Goldringer, I., Mendes-Moreira, P., & Chable, V. Lacape, J. M. (2008). Use of molecular markers in participatory plant
(2012). Analysis of major PPB experiences worldwide. Retrieved from http:// breeding: Assessing the genetic variability in cotton populations bred
www.solibam.eu/modules/wfdownloads/singlefile.php?cid=12&lid=33 by farmers. Annals of Applied Biology, 152, 113–119.
Ceccarelli, S., Galie, A., & Grando, S. (2013). Participatory breeding for cli- Lawrence, C. J., Harper, L. C., Schaeffer, M. L., Sen, T. Z., Seigfried, T. E., &
mate change-related traits. In C. Kole (Ed.), Genomics and breeding for Campbell, D. A. (2008). MaizeGDB: The maize model organism data-
climate resilient crops. Concepts and strategies. Volume 1 (pp. 331–376). base for basic, translational, and applied research. International Journal
Heidelberg, Germany: Springer. of Plant Genomics, 2008, 496957.
Ceccarelli, S., Grando, S., Maatougui, M., Michae, L. M., Slash, M., Lipka, A. E., Gore, M. A., Magallanes-Lundback, M., Mesberg, A., Lin, H.,
Haghparast, R., … Nachit, M. (2010). Plant breeding and climate Tiede, T., … DellaPenna, D. (2013). Genome-­wide association study
changes. Journal of Agricultural Science, 148, 627–637. and pathway-­level analysis of tocochromanol levels in maize grain. G3:
Cleveland, D. A., Soleri, D., & Smith, S. E. (1999). Farmer plant breed- Genes|Genomes|Genetics, 3, 1287–1299.
ing from biological perspective: Implications for collaborative plant Liu, R. H. (2003). Health benefits of fruit and vegetables are from additive
breeding. CIMMYT Economics Working Paper CIMMYT, Mexico DF: and synergistic combinations of phytochemicals. American Journal of
CIMMYT, 99–10. Clinical Nutrition, 78, 517S–520S.
Do, C., Waples, R. S., Peel, D., Macbeth, G. M., Tillett, B. J., & Ovenden, J. R. Lopez-Martinez, L. X., Oliart-Ros, R. M., Valerio-Alfaro, G., Lee, C. H., Parkin,
(2014). NeEstimator v2: Re-­implementation of software for the estima- K. L., & Garcia, H. S. (2009). Antioxidant activity, phenolic compounds
tion of contemporary effective population size (Ne) from genetic data. and anthocyanins content of eighteen strains of Mexican maize. Food
Molecular Ecology Resources, 14, 209–214. Science and Technology, 42, 1187–1192.
Excoffier, L., Laval, G., & Schneider, S. (2005). Arlequin ver. 3.0: An in- Mendes-Moreira, P. M. R., Pego, S. E., Vaz Patto, C., & Hallauer, A. R. (2008).
tegrated software package for population genetics data analysis. Comparison of selection methods on ‘Pigarro’, a Portuguese improved
Evolutionary Bioinformatics Online, 1, 47–50. maize population with fasciation expression. Euphytica, 163, 481–499.
ALVES et al. |
      17

Mendes-Moreira, P. M. R., Vaz Patto, M. C., Mota, M., Mendes-Moreira, Silva, S., Gomes, L., Leitão, F., Coelho, A. V., & Vilas Boas, L. (2006). Phenolic
J., Santos, J. P. N., Santos, J. P. P., … Pego, S. E. (2009). ‘Fandango’: compounds and antioxidant activity of Olea europaea L. fruits and
Long term adaptation of exotic germplasm to a Portuguese on-­farm-­ leaves. Food Science Technology International, 12, 385–396.
conservation and breeding project. Maydica, 54, 269–285. Singleton, V. L., Orthofer, R., & Lamuela-Raventos, R. M. (1999). Analysis of
Murphy, K. M., Carter, A. H., & Jones, S. S. (2013). Evolutionary breeding total phenols and other oxidant substrates and antioxidants by means
and climate change. In C. Kole (Ed.), Genomics and breeding for cli- of Folin-­Ciocalteu reagent. Methods in Enzymology, 299, 157–178.
mate resilient crops. Concepts and strategies. Volume 1 (pp. 377–389). Smith, M. E., Castillo, F. G., & Gómez, F. (2001). Participatory plant breeding
Heidelberg, Germany: Springer. with maize in Mexico and Honduras. Euphytica, 122, 551–565.
Nei, M., & Tajima, F. (1981). Genetic drift and estimation of effective popu- Solomon, K. F., Martin, I., & Zeppa, A. (2010). Temporal genetic structure
lation size. Genetics, 98, 625–640. patterns in tropical maize populations under reciprocal recurrent selec-
Ortiz, R. (2011). Agrobiodiversity management for climate change. In J. M. tion. Euphytica, 176, 239–249.
Lenné, & D. Wood (Eds.), Agrobiodiversity management for food security Vaz Patto, M. C., Mendes-Moreira, P. M., Alves, M. L., Mecha, E., Brites,
(pp. 189–211). Wallingford, Oxon, United Kingdom: CAB International. C., Bronze, R., & Pego, S. (2013). Participatory plant quality breeding:
Owens, B. F., Lipka, A. E., Magallanes-Lundback, M., Tiede, T., Diepenbrock, An ancient art revisited by knowledge sharing. The Portuguese expe-
C. H., Kandianis, C. B., … Rocheford, T. (2014). A foundation for provita- rience. In S. B. Andersen (Ed.), Plant breeding from laboratories to fields
min A biofortification of maize: Genome-­wide association and genomic (pp. 255–288). Rijeka: InTech.
prediction models of carotenoid levels. Genetics, 198, 1699–1716. Vaz Patto, M. C., Moreira, P. M., Almeida, N., Satovic, Z., & Pego, S. (2008).
Phillips, S. L., & Wolfe, M. S. (2005). Evolutionary plant breeding for low Genetic diversity evolution through participatory maize breeding in
input systems. The Journal of Agricultural Science, 143, 245–254. Portugal. Euphytica, 161, 283–291.
Pingali, P. L. (ed.). (2001). CIMMYT 1999–2000 World Maize Facts and Vaz Patto, M. C., Moreira, P. M., Carvalho, V., & Pego, S. (2007). Collecting
Trends. Meeting World Maize Needs: Technological Opportunities and maize (Zea mays L. convar. mays) with potential technological ability
Priorities for the Public Sector Mexico. D.F.: CIMMYT. for bread making in Portugal. Genetic Resources and Crop Evolution, 54,
Rauf, S., Teixeira da Silva, J. A., Khan, A. A., & Naveed, A. (2010). 1555–1563.
Consequences of plant breeding on genetic diversity. International Waples, R. S. (1989a). Temporal variation in allele frequencies: Test the right
Journal of Plant Breeding, 4, 1–21. hypothesis. Evolution, 43, 1236–1251.
Raymond, M., & Rousset, F. (1995). GENEPOP (version 1.2): Population ge- Waples, R. S. (1989b). A generalized approach for estimating effective pop-
netics software for exact tests and ecumenicism. Journal of Heredity, ulation size from temporal changes in allele frequency. Genetics, 95,
86, 248–249. 489–502.
Rether, B., Delmas, G., & Laouedj, A. (1993). Isolation of polysaccharide-­ Wheeler, T., & von Braun, J. (2013). Climate change impacts on global food
free DNA from plants. Plant Molecular Biology Reporter, 11, 333–337. security. Science, 341, 508–513.
Rice, W. R. (1989). Analyzing tables of statistical tests. Evolution, 43,
223–225.
Saghai-Maroof, M. A., Soliman, K. M., Jorgensen, R. A., & Allard, R. W. S U P P O RT I NG I NFO R M AT I O N
(1984). Ribosomal DNA spacer-­length polymorphisms in barley:
Mendelian inheritance, chromosomal location, and population dynam- Additional Supporting Information may be found online in the
ics. Proceedings of the National Academy of Sciences of the United States ­supporting information tab for this article. 
of America, 81, 8014–8018.
SAS Institute Inc (2004). The SAS system for windows, V8. 02. Cary, NC: SAS
Institute.
How to cite this article: Alves ML, Belo M, Carbas B, et al.
Schuelke, M. (2000). An economic method for the fluorescent labeling of
PCR fragments. Nature Biotechnology, 18, 233–234. Long-­term on-­farm participatory maize breeding by stratified
Silva, S., Bronze, M. R., Figueira, M. E., Siwy, J., Mischak, H., Combet, E., mass selection retains molecular diversity while improving
& Mullen, W. (2015). Impact of a 6-­wk olive oil supplementation in agronomic performance. Evol Appl. 2017;00:1–17.
healthy adults on urinary proteomic biomarkers of coronary artery
https://doi.org/10.1111/eva.12549
disease, chronic kidney disease, and diabetes (types 1 and 2): A ran-
domized, parallel, controlled, double-­blind study. The American Journal
of Clinical Nutrition, 101, 44–54.