**3. Participatory plant quality breeding in Portugal**

Presently, in particular Portuguese regions, known by their high quality maize bread, farm‐ ers keep on cultivating their traditional landraces. Traditionally selected landraces are main‐ ly white kernel flint types, demonstrating quality over yield and maintaining genetic diversity to increase adaptability to a large variety of edaphic/climatic conditions, such as drought or aluminium toxicity [37].

Maize is definitely a deep-rooted crop in the Portuguese rural tradition and the available ge‐ netic variability of its landraces offers a superb challenge for breeding for special quality traits. Since the on going PPB project at Sousa Valley (VASO) revealed promising breeding results, our objective is to get further support to maintain the actual project and extend it to other maize landraces production areas as a way to increase the use value of this traditional germplasm and by doing so promote *in-situ/on-farm* conservation and halt the serious genet‐ ic diversity erosion. The inclusion of organic breeding objectives in the actual PPB project is also being considered to add value to these improved landraces since the low input sustain‐ able farming system typical of the traditional production systems is already very similar to the organic farming directives. In this more recent PPB approach, molecular and detailed quality data will be used in order to increase the effectiveness of selection when appropriate.

#### **3.1. New material and farmers prospection**

With the idea of establishing an on-farm conservation project, with farmers' engagement through participatory breeding approaches, to halt the genetic erosion by improving those landraces and increasing their market value, members of our team engaged in a field expe‐ dition to the Central region of Portugal to collect enduring landraces [7]. In this region farm‐ ers grow maize landraces, known for their good maize bread quality, in association with common bean local varieties, in a traditional intercropping practice. The collected landraces represent important sources of genes and gene combinations not yet available for crop qual‐ ity breeding programs and due to their intrinsic quality traits (that promoted their mainte‐ nance in cultivation) are the best candidates for expanding the already existing participatory breeding program (VASO) to other regions with more emphasis on quality breeding. Around 50 different (yellow and white) maize landraces were collected, characterized using pre-breeding approaches and conserved in cold storage [7]. These landraces, together with other landraces that were subsequently collected from the surrounding regions, represented the basis for the PPB net existing in the country. During this expedition several associated crop (beans, rye and pumpkin) landraces were also collected. The collected bean landraces are also now being characterized at agronomical, genetic diversity and quality level to select the best material to implement also participatory breeding approaches.

During this expedition, the first steps on contacting new farmers to extend the original PPB net were also given and the most enthusiastic farmers meet during those days, are now in‐ volved in the participatory research. VASO still continues nowadays and it is the best inspi‐ ration for those intending to start their PPB programs. For this reason, VASO initial actors have been invited to give their testimony to the new associated farmers within the actual ex‐ tended PPB net. This action boosted the program giving new perspectives to the new farm‐ ers, and new paths to our program. Some meetings have been prepared to have farmers' perception (e.g. know what are the kernel preferences for maize bread), or where research‐ ers present their achievements to the network, with time for discussion (e.g. soil, agronomic traits, genetic diversity and quality). Along this process an identification of the farmers' pro‐ file was made, their motivations and interests (e.g. germplasm development, trials) in order to fully engage them with the project. Even though the majority of farmers has no back‐ ground on basic statistics (e.g. replication), they do their empirical research. For this reason appreciate the contact with a "practical" academia, where both look in the same direction and where the arena is at their farmers fields. This action demands an intensive networking regarding motivation and science, where future perspectives are discussed, and results de‐ part from farmers' fields.

#### **3.2. Quality breeding objectives**

germplasm and by doing so promote *in-situ/on-farm* conservation and halt the serious genet‐ ic diversity erosion. The inclusion of organic breeding objectives in the actual PPB project is also being considered to add value to these improved landraces since the low input sustain‐ able farming system typical of the traditional production systems is already very similar to the organic farming directives. In this more recent PPB approach, molecular and detailed quality data will be used in order to increase the effectiveness of selection when appropriate.

With the idea of establishing an on-farm conservation project, with farmers' engagement through participatory breeding approaches, to halt the genetic erosion by improving those landraces and increasing their market value, members of our team engaged in a field expe‐ dition to the Central region of Portugal to collect enduring landraces [7]. In this region farm‐ ers grow maize landraces, known for their good maize bread quality, in association with common bean local varieties, in a traditional intercropping practice. The collected landraces represent important sources of genes and gene combinations not yet available for crop qual‐ ity breeding programs and due to their intrinsic quality traits (that promoted their mainte‐ nance in cultivation) are the best candidates for expanding the already existing participatory breeding program (VASO) to other regions with more emphasis on quality breeding. Around 50 different (yellow and white) maize landraces were collected, characterized using pre-breeding approaches and conserved in cold storage [7]. These landraces, together with other landraces that were subsequently collected from the surrounding regions, represented the basis for the PPB net existing in the country. During this expedition several associated crop (beans, rye and pumpkin) landraces were also collected. The collected bean landraces are also now being characterized at agronomical, genetic diversity and quality level to select

During this expedition, the first steps on contacting new farmers to extend the original PPB net were also given and the most enthusiastic farmers meet during those days, are now in‐ volved in the participatory research. VASO still continues nowadays and it is the best inspi‐ ration for those intending to start their PPB programs. For this reason, VASO initial actors have been invited to give their testimony to the new associated farmers within the actual ex‐ tended PPB net. This action boosted the program giving new perspectives to the new farm‐ ers, and new paths to our program. Some meetings have been prepared to have farmers' perception (e.g. know what are the kernel preferences for maize bread), or where research‐ ers present their achievements to the network, with time for discussion (e.g. soil, agronomic traits, genetic diversity and quality). Along this process an identification of the farmers' pro‐ file was made, their motivations and interests (e.g. germplasm development, trials) in order to fully engage them with the project. Even though the majority of farmers has no back‐ ground on basic statistics (e.g. replication), they do their empirical research. For this reason appreciate the contact with a "practical" academia, where both look in the same direction and where the arena is at their farmers fields. This action demands an intensive networking regarding motivation and science, where future perspectives are discussed, and results de‐

the best material to implement also participatory breeding approaches.

**3.1. New material and farmers prospection**

268 Plant Breeding from Laboratories to Fields

part from farmers' fields.

In the national Portuguese panorama we are now more concerned with the landraces still in cultivation and the quality that prevented their replacement by the generalized hybrid use. These materials represent the actual enduring and surviving genotypes and so constitute the best candidates for a quality breeding program. Additionally, these materials should be al‐ lowed to pursue a natural evolution under *in situ* conservation and farmers must be reward‐ ed for their contribution to halt the current and continuing loss of plant diversity. The only way to achieve this is to promote a sustainable use of plant diversity, where conservation responsibilities and benefits will be shared with farmers [7]. A participatory approach seems to be the most logical solution. We try to identify ways of supporting farmers in the mainte‐ nance of traditional varieties and crop genetic diversity by performing better with their own seeds improved.

Portuguese maize landraces have been preserved on-farm, due to particular quality traits not found on their competing modern hybrid varieties. These landraces are mainly flint type open pollinated varieties (OPV) with technological ability for the production of the tradi‐ tional maize leavened bread called "broa" that still plays an important economic and social role on Central and Northern rural communities of the country [7].

Due to this, we decided to start studying in more detail the technological ability for bread production of maize landraces as the major quality trait to breed for. However, later on, it has been described that other quality traits, such a flavor or aromas were also contributing to the consumers preferences for bread obtained from traditional maize landraces in detri‐ ment of maize hybrid varieties bread [43]. Volatile components responsible for the aroma were then also included into our detailed study. Presently and due to consumers higher con‐ cern about the quality of their food and how their diet can influence their well being also antioxidant compounds and their bioactivity are also being analyzed on the flour and bread made from our traditional maize varieties.

Our objective is that our improved varieties will be attractive to consumers, processing in‐ dustry and farmers, answering health and environmental public concerns and increasing sustainability of farming systems.

#### *3.2.1. Technological ability for bread production*

The introduction of maize in the Iberian Peninsula during the fifteenth century produced important changes in agriculture and in the diet of the people. Maize has become highly in‐ tegrated into Portuguese agriculture and diet, and and appears as the major cereal for bread making in the middle of 19th century. The bread produced at that time was "broa" where maize flour meal was mixed with wheat or rye. Presently, Portuguese rural areas continue to produce "broa", mainly in the northwest parts of the country in a wide variety of recipes, some with protected geographical identification and traditional methods of baking. The quality of "broa" is the result of empirical knowledge that is very closely related to the qual‐ ity of the maize, kernel processing, blending flours and baking procedures, including fer‐ mentation and baking.

There are many traditional recipes to prepare "broa", but the traditional process involves adding maize flour (sieved whole meal flour ranging between 50% and 80%), hot water, wheat and rye flours, yeast and leavened dough from the late "broa" (acting as sourdough). After mixing, resting and proofing, the dough is baked in a wood-fired oven. This empirical process leads to an ethnic product highly accepted for its distinctive sensory characteristics.

The process begins with blanching maize flour in water boiled followed by kneading. Blanch‐ ing is important to obtain high consistency dough's because in the absence or reduced amount of gluten the dough rheological properties are provided by the starch gelatinization [43]. The addition of sourdough is another important aspect in the preparation of "broa", and its microbial diversity has been characterized [44] with mainly lactic acid bacteria *Lactobacillus* (*brevis, bulgaris plantarum*) being present in addition to yeast (*Saccharomyces cerevisiae*).

In terms of maize physical characteristics, kernel size and shape, weight and density, degree of stress cracks and resistance to milling and compression have all been linked to hardness and this is the primary cause for the large differences in rheological properties of flours which have subsequent effects on processing.

The majority of commercial hybrid maize varieties that are currently grown in Portugal are dent type, but also some flour types can be found, characterized by a soft or floury endo‐ sperm respectively. However, in the traditional varieties and landraces predominates the flint type with the hardest kernels, resulting from the presence of a large and continuous volume of horny (vitreous) endosperm. Maize flint type has harder endosperm than dent types, resulting in different viscosity profiles. Flours from maize flint grain have lower peak viscosity and lower retrogradation than dent types [45, 46].

White maize is the preferred choice by northwest rural populations, probably due to cultur‐ al and historical reasons that have created food habits. Indeed in the 18th century, when the maize was the main cereal used for bread making, white bread was the most appreciated, symbolizing wealth and prestige [47], in this context white maize flour was the most suita‐ ble for blending with wheat flours. However, it is important to understand if there are dif‐ ferences between the rheological behavior of maize flour from white and yellow grains. With this aim, we analyzed a collection of maize OPVs and found no significant correlation between the Colour Chromameter b\* - yellow/blue index and viscosity parameters of flours [34]. From the nutritional standpoint, the white grain has the disadvantage that it is devoid of carotenoids, which are important antioxidants for health. Nevertheless, our preliminary results indicate high amounts of other nutritional compounds such as tocopherols.

Kernel processing into milling is an important quality factor for the production of "broa" be‐ cause it determines the performance of the flour. Dry-milling process is used for the produc‐ tion of maize meal used for bread making and whole grain is processed traditionally in stone wheel mills, moved by water or wind, and nowadays frequently by electricity.

The native starch can be damaged to a greater or lesser extent thereby influencing the flour water absorption capacity and enzymatic attack, especially α-amylase. The type of grinding may also affect ash flour content, which interferes with the evolution of pH during the fer‐ mentation step. The grinding mills driven by water occurs at a slower rate, flours obtained with this process have lower ash content, lower proportion of damaged starch, and higher maximum viscosity than obtained in electrical mills [46].

In addition to the "broa" sensory specificities and the need to diversify baking products to fulfill the consumers' appreciation range of traditional breads, there are also reasons related to nutrition disorders that promote the study of maize quality for bread making.

The high indices of chronic diseases, such as obesity and diabetes, increase the demand for the development of breads with starch that is slowly digestible or partially resistant to the digestive process namely resistant starch [48]. "Broa" revealed a greater resistant starch con‐ tent than the wheat bread [49]. Differences in starch digestibility or type of dietary fiber, the typical fermentation and bread volume also contributes for lower glycemic index of "broa" when compared with wheat bread [49]. Gluten enteropathy (coeliac disease) is another seri‐ ous chronic disease, caused by an inappropriate immune response to dietary wheat gluten or similar proteins of barley or rye. Maize is a gluten-free cereal, thus suitable to produce foods addressed to celiac patients. The acquired knowledge on "broa" (made from compo‐ site maize–rye–wheat flour) is important for facing the challenges of producing gluten-free bread that usually exhibits compact crumb texture and low specific volume [43]. Baking as‐ says were performed and demonstrated that bread making technology could be satisfactori‐ ly applied to produce gluten-free "broa" [43].

Strategies to further improve maize kernel quality for "broa" production considering flour rheological properties and nutrients are under intense investigation, mainly focused on vis‐ cosity profiles, protein content, carotenoids and tocopherols.

Management of large number of accessions implies adoption of rapid and non destructive tests for efficiently screening quality traits, consequently research on Near Infrared Spectro‐ scopy (NIR) models to estimate maize kernels quality is under progress and it will be of ex‐ treme importance as a fast and inexpensive way to support quality PPB.

Moreover, the selection parameters adopted in quality PPB should reflect "broa" consumers' preference and therefore "broa" bread sensory analyzes with consumer panel are being im‐ plemented and the data obtained will be used to improve the screening quality maize tests for bread making.

#### *3.2.2.Aroma*

There are many traditional recipes to prepare "broa", but the traditional process involves adding maize flour (sieved whole meal flour ranging between 50% and 80%), hot water, wheat and rye flours, yeast and leavened dough from the late "broa" (acting as sourdough). After mixing, resting and proofing, the dough is baked in a wood-fired oven. This empirical process leads to an ethnic product highly accepted for its distinctive sensory characteristics. The process begins with blanching maize flour in water boiled followed by kneading. Blanch‐ ing is important to obtain high consistency dough's because in the absence or reduced amount of gluten the dough rheological properties are provided by the starch gelatinization [43]. The addition of sourdough is another important aspect in the preparation of "broa", and its microbial diversity has been characterized [44] with mainly lactic acid bacteria *Lactobacillus* (*brevis, bulgaris plantarum*) being present in addition to yeast (*Saccharomyces cerevisiae*).

In terms of maize physical characteristics, kernel size and shape, weight and density, degree of stress cracks and resistance to milling and compression have all been linked to hardness and this is the primary cause for the large differences in rheological properties of flours

The majority of commercial hybrid maize varieties that are currently grown in Portugal are dent type, but also some flour types can be found, characterized by a soft or floury endo‐ sperm respectively. However, in the traditional varieties and landraces predominates the flint type with the hardest kernels, resulting from the presence of a large and continuous volume of horny (vitreous) endosperm. Maize flint type has harder endosperm than dent types, resulting in different viscosity profiles. Flours from maize flint grain have lower peak

White maize is the preferred choice by northwest rural populations, probably due to cultur‐ al and historical reasons that have created food habits. Indeed in the 18th century, when the maize was the main cereal used for bread making, white bread was the most appreciated, symbolizing wealth and prestige [47], in this context white maize flour was the most suita‐ ble for blending with wheat flours. However, it is important to understand if there are dif‐ ferences between the rheological behavior of maize flour from white and yellow grains. With this aim, we analyzed a collection of maize OPVs and found no significant correlation between the Colour Chromameter b\* - yellow/blue index and viscosity parameters of flours [34]. From the nutritional standpoint, the white grain has the disadvantage that it is devoid of carotenoids, which are important antioxidants for health. Nevertheless, our preliminary

results indicate high amounts of other nutritional compounds such as tocopherols.

stone wheel mills, moved by water or wind, and nowadays frequently by electricity.

Kernel processing into milling is an important quality factor for the production of "broa" be‐ cause it determines the performance of the flour. Dry-milling process is used for the produc‐ tion of maize meal used for bread making and whole grain is processed traditionally in

The native starch can be damaged to a greater or lesser extent thereby influencing the flour water absorption capacity and enzymatic attack, especially α-amylase. The type of grinding may also affect ash flour content, which interferes with the evolution of pH during the fer‐ mentation step. The grinding mills driven by water occurs at a slower rate, flours obtained

which have subsequent effects on processing.

270 Plant Breeding from Laboratories to Fields

viscosity and lower retrogradation than dent types [45, 46].

Aroma strongly influences food quality and therefore consumer's preferences and accepta‐ bility for the products. Sweet, sour, salty, bitter and umami tastes, olfactory responses, oral sensory sensations related with astringency, coolness and pressure, contribute to food aro‐ ma [50]. Since olfactory responses involve a huge number of descriptors to distinguish hun‐ dreds of different odors, it is not surprising that most of the work developed in aroma research has been related with volatile compounds analysis [51].

Volatile compounds in plants and foods are produced during harvesting and processing, by enzymatic degradation [52]. The type of volatile compounds depends on plant species, gen‐ otype, plant part and environmental growing conditions [53]. Alcohols, aldehydes, ke‐ tones, hydrocarbons and terpenic compounds are among the main volatile compounds responsible for foods' aroma and most of them are present in trace amounts, which diffi‐ cult the task of aroma analysis [50]. Drying, handling, milling and storage conditions may affect aroma of foods [50].

Solid Phase Micro Extraction (SPME) combined with Gas Chromatography-Mass Spectrom‐ etry (GC-MS) is now the most used technique for the analysis of volatile compounds [52]. Aroma volatile compounds of different maize types and preparations have been studied by SPME-GC-MS. Characteristic odor of dimethyl sulfide has been associated with sweet maize aroma. Other important compounds include 1-hydroxy-2-propanone, 2-hydroxy-3-butanone and 2,3-butanediol. Higher concentrations of such volatile compounds were reported in can‐ ned maize, when comparing canned, frozen and fresh maize [50, 54]. In popcorn, 6-acetylte‐ trahydropyridine, 2-acetyl-1-pyrroline and 2-propionyl-1-pirroline were described as the most important aroma compounds [54]. In maize tortilla and taco shell it was possible to identify an aroma component not previously identified, 2-aminoacetophenone [54].

Information about aroma volatile compounds in Portuguese maize and maize bread, until now, is scarce. In order to characterize these compounds in this national germplasm and re‐ spective food products (bread), we studied a collection of 51 Portuguese maize landraces representing the starting material of the present national participatory maize breeding project. Solid Phase Micro Extraction (SPME) combined with gas chromatography and mass spectrometry was used for the analysis. Volatiles were identified based on comparison with mass spectra in reference libraries NIST 21.LIB and Willey 229.LIB and by the Linear Reten‐ tion Index (LRI). Aldehydes (hexanal, heptanal, nonanal, 2-nonenal (E), and decanal) were identified as main volatile compounds responsible for maize flour aroma being Hexanal the most representative aldehyde compound on the analyzed flour.

The analysis of the aroma volatile compounds released from traditional Portuguese bread ("broa") made from selected maize varieties is under way using the same conditions of analysis.

#### *3.2.2. Phenolic compounds*

Phenolic compounds are secondary metabolites produced by plants as protection against fungi, herbivores, UV radiation and oxidative cell injury, revealing also important functions in several aspects of plant life as growth, pigmentation and reproduction [55].

Phenolic compounds can contribute, with other dietary components such as vitamins C, E and carotenoids to the human protection against oxidative stress caused by an excess of re‐ active oxygen species (ROS) [55, 56]. Their antioxidant activity contributes to the inhibition of oxidative mechanisms underlying several degenerative diseases such as diabetes, cardio‐ vascular diseases, and cancer [57, 58]. Besides their health promoting effect phenolic acids present in maize samples, for example, may contribute indirectly to flavor quality trough in‐ hibition of lipid oxidation [50].

Phenolic compounds can be classified into two major classes, flavonoids and non-flavo‐ noids. Phenolic acids (hydroxybenzoic acids, hydroxycinnamic acids) and flavonoids corre‐ spond to soluble compounds (easily extracted with polar solvents such as ethanol, methanol and mixtures with water) which can be separated and identified by High Performance Liq‐ uid Chromatography (HPLC) and detected in the UV-Vis and by Mass Spectrometry (MS) [59].

tones, hydrocarbons and terpenic compounds are among the main volatile compounds responsible for foods' aroma and most of them are present in trace amounts, which diffi‐ cult the task of aroma analysis [50]. Drying, handling, milling and storage conditions may

Solid Phase Micro Extraction (SPME) combined with Gas Chromatography-Mass Spectrom‐ etry (GC-MS) is now the most used technique for the analysis of volatile compounds [52]. Aroma volatile compounds of different maize types and preparations have been studied by SPME-GC-MS. Characteristic odor of dimethyl sulfide has been associated with sweet maize aroma. Other important compounds include 1-hydroxy-2-propanone, 2-hydroxy-3-butanone and 2,3-butanediol. Higher concentrations of such volatile compounds were reported in can‐ ned maize, when comparing canned, frozen and fresh maize [50, 54]. In popcorn, 6-acetylte‐ trahydropyridine, 2-acetyl-1-pyrroline and 2-propionyl-1-pirroline were described as the most important aroma compounds [54]. In maize tortilla and taco shell it was possible to

identify an aroma component not previously identified, 2-aminoacetophenone [54].

most representative aldehyde compound on the analyzed flour.

Information about aroma volatile compounds in Portuguese maize and maize bread, until now, is scarce. In order to characterize these compounds in this national germplasm and re‐ spective food products (bread), we studied a collection of 51 Portuguese maize landraces representing the starting material of the present national participatory maize breeding project. Solid Phase Micro Extraction (SPME) combined with gas chromatography and mass spectrometry was used for the analysis. Volatiles were identified based on comparison with mass spectra in reference libraries NIST 21.LIB and Willey 229.LIB and by the Linear Reten‐ tion Index (LRI). Aldehydes (hexanal, heptanal, nonanal, 2-nonenal (E), and decanal) were identified as main volatile compounds responsible for maize flour aroma being Hexanal the

The analysis of the aroma volatile compounds released from traditional Portuguese bread ("broa") made from selected maize varieties is under way using the same conditions of analysis.

Phenolic compounds are secondary metabolites produced by plants as protection against fungi, herbivores, UV radiation and oxidative cell injury, revealing also important functions

Phenolic compounds can contribute, with other dietary components such as vitamins C, E and carotenoids to the human protection against oxidative stress caused by an excess of re‐ active oxygen species (ROS) [55, 56]. Their antioxidant activity contributes to the inhibition of oxidative mechanisms underlying several degenerative diseases such as diabetes, cardio‐ vascular diseases, and cancer [57, 58]. Besides their health promoting effect phenolic acids present in maize samples, for example, may contribute indirectly to flavor quality trough in‐

Phenolic compounds can be classified into two major classes, flavonoids and non-flavo‐ noids. Phenolic acids (hydroxybenzoic acids, hydroxycinnamic acids) and flavonoids corre‐ spond to soluble compounds (easily extracted with polar solvents such as ethanol, methanol

in several aspects of plant life as growth, pigmentation and reproduction [55].

affect aroma of foods [50].

272 Plant Breeding from Laboratories to Fields

*3.2.2. Phenolic compounds*

hibition of lipid oxidation [50].

Some studies have been conducted in order to characterize maize polyphenolic content, and vanillic, *p*-coumaric, ferulic, protocatechuic acids, derivatives of hesperitin, quercetin and anthocyanins like cyanidin-3-glucoside and pelargonidin-3-glucoside [60] were identified as the most important ones. The actual knowledge about phenolic compounds bioaccessibility and bioavailability [61], contributing to the protective effect in biological systems, is still scarce.

In common beans (*Phaseolus vulgaris* L.) phenolic compounds (phenolic acids and flavo‐ noids) were mostly described in the seed coat and at lower amounts in cotyledons [58, 62, 63]. Compounds, such as *p*-hydroxybenzoic, vanillic, caffeic, syringic, coumaric, ferulic and synapic acids as well as flavonoids such as quercetin, kaempferol, daidzein, genistein, *p*-cou‐ mestrol and anthocyanins like delphinidin and cyaniding were already identified in com‐ mon beans [63]. It is widely accepted that thermal processing (boiling or steaming treatment) affects phenolic compounds content and antioxidant activity values [64, 65].

In relation to the Portuguese maize and beans germplasm, and to our knowledge, no infor‐ mation on the phenolic compounds as ever been published. So our initial goal was to char‐ acterize the flour composition of 51 Portuguese maize landraces and of 32 different varieties of Portuguese beans.

Spectrophotometric assays were performed to determine total phenolic and total flavonoids content in the samples. For maize, total phenolic content ranged from 100.30 ± 4.81 to 206.83 ± 9.55 mg of gallic acid equivalents/ 100g DW (dry weight) and total flavonoids content ranged between 0.69 ± 0.07 and 17.01 ± 0.52 mg of catechin equivalents/ 100g DW. For beans, the total phenolic content ranged between 1.00 ± 0.02 and 6.83 ± 0.31 mg of gallic acid equiv‐ alents/g and total flavonoids content ranged between 0.09 ± 0.00 and 2.50 ± 0.01 mg of cate‐ chin equivalents/g.

With the main objective of identifying soluble free, soluble conjugated and insoluble phenol‐ ic compounds in maize flour by HPLC, acidic and alkaline hydrolysis were performed. The phenolic fraction which presented higher amount of compounds corresponded to the in‐ soluble. Using HPLC with diode array detector (DAD) it was possible to identify *p*-coumaric and ferulic acids as well as aldehydes such as vanillin and syringaldehyde. In bean's ex‐ tracts, phenolic acids such as caffeic acid and flavonoids such as catechin, quercetin-3-O-ru‐ tinosideand kaempferol-3-O-glucoside were identified.

Studies of the antioxidant activity by Oxygen Radical Absorbance Capacity (ORAC) have also started and were already performed for some maize flour extracts. Values obtained range from 364.30-1223.55 μmol Trolox Equivalents Antioxidant Capacity (TEAC)/100g. In bean extracts, values obtained were between 28.99 ± 2.09 and 189.12 ± 10.20 μmol TEAC/g. These ORAC values showed a strong positive correlation with total phenolic content (R=0.9087) and total flavonoids content (R=0.9171), evaluated by colorimetric methods. The results obtained, until now, revealed a great variability of polyphenolic content and antioxidant activity in the samples analyzed anticipating a high potential for quality breeding within these materials.

Future studies for phenolic compounds' identification and quantification by HPLC-DAD and LC-MS/MS will be performed in raw and processed maize, whole beans seeds and beans fractions (seed coat and cotyledons obtained after beans soaking) submitted to acidic, enzymatic and alkaline hydrolysis. Those studies will allow recognition of the digestion im‐ pact on maize and beans' phenolic composition and represent very interesting information to provide to the consumer. This information may increase the crops market value and should be taken into consideration on future participatory breeding selection.

#### **3.3. Development of molecular tools for assisting quality selection**

Genetics, particularly molecular genetics, provides further information on patterns of diver‐ sity distribution and allows the investigation of the relation of observed diversity with envi‐ ronment, social and cultural factors, providing means to reconcile farmer's classification schemes with genetic distinctiveness. It also helps determine whether there is a wide enough genetic base for future improvement of the *in-situ* materials, or whether there is suf‐ ficient diversity to provide system resilience [6]. It can also underpin the identification of ways of supporting the maintenance of traditional varieties, such as in supporting protected geographical identification of certain plant or crop product.

Presently, in our extended PPB project we are conjugating the identification of agronomic and specific quality traits with molecular characterization so as to exploit efficiently the local diversity and produce varieties that are superior in marginal environments, but have a broad genetic base and a high quality level. Nevertheless, in Portugal, molecular breeding is still given its first steps.

In this section we will summarize the development of molecular tools to assist the imple‐ mentation of participatory breeding program focusing on maize improvement for produc‐ ing high quality bread. One of the key elements for the implementation of a successful breeding program is the existence of decision supporting tools. Different molecular markers are being tested in order to create new decision supporting tools. Among the different classes of molecular markers, we started initially to use simple sequence repeat (SSR or mi‐ crosatellite) markers that have proven to be the marker of choice for a variety of applica‐ tions, particularly in breeding [66]. We are now starting to use also single nucleotide polymorphisms (SNPs) molecular markers that are more abundant in the genome and ame‐ nable to automation for high-throughput genotyping [67].

Molecular markers are being used to achieve two main research objectives. First we are us‐ ing molecular markers to evaluate the progress obtained in conserving or increasing diversi‐ ty through participatory breeding, as already described in the section 2.3 of PPB success evaluation [42]. The genetic diversity of the newly introduced maize and beans landraces in‐ to the participatory plant breeding net is now also routinely characterized, with 20 to 22 SSR uniformly distributed throughout the maize and bean genomes respectively. This method enables us to check if sufficient diversity is present to allow selection and to select the most promising landraces in order to increase the genetic diversity by crossing genetically distant landraces. These studies have also allowed us to compare the genetic diversity with quality clustering of landraces [34]. In detail, 46 traditional maize landraces collected from known high quality maize bread Portuguese regions, plus six participatory improved maize OPVs from the VASO project, were analyzed for eight different parameters related with their tech‐ nological ability for bread production, and 13 SSR markers. It was possible to classify these OPVs into three distinct clusters based on the quality traits. Nevertheless, no clear clustering based on genetic distances was observed despite the high levels of genetic diversity present‐ ed by these Portuguese landraces [34]. Based on the existence of diversity at molecular level and high quality, the Portuguese maize landraces conserved on-farm represent valuable germplasm with high potential for bread quality improvement [34]. This study also provid‐ ed important information for the selection of landraces to keep under the extended PPB project.

Future studies for phenolic compounds' identification and quantification by HPLC-DAD and LC-MS/MS will be performed in raw and processed maize, whole beans seeds and beans fractions (seed coat and cotyledons obtained after beans soaking) submitted to acidic, enzymatic and alkaline hydrolysis. Those studies will allow recognition of the digestion im‐ pact on maize and beans' phenolic composition and represent very interesting information to provide to the consumer. This information may increase the crops market value and

Genetics, particularly molecular genetics, provides further information on patterns of diver‐ sity distribution and allows the investigation of the relation of observed diversity with envi‐ ronment, social and cultural factors, providing means to reconcile farmer's classification schemes with genetic distinctiveness. It also helps determine whether there is a wide enough genetic base for future improvement of the *in-situ* materials, or whether there is suf‐ ficient diversity to provide system resilience [6]. It can also underpin the identification of ways of supporting the maintenance of traditional varieties, such as in supporting protected

Presently, in our extended PPB project we are conjugating the identification of agronomic and specific quality traits with molecular characterization so as to exploit efficiently the local diversity and produce varieties that are superior in marginal environments, but have a broad genetic base and a high quality level. Nevertheless, in Portugal, molecular breeding is

In this section we will summarize the development of molecular tools to assist the imple‐ mentation of participatory breeding program focusing on maize improvement for produc‐ ing high quality bread. One of the key elements for the implementation of a successful breeding program is the existence of decision supporting tools. Different molecular markers are being tested in order to create new decision supporting tools. Among the different classes of molecular markers, we started initially to use simple sequence repeat (SSR or mi‐ crosatellite) markers that have proven to be the marker of choice for a variety of applica‐ tions, particularly in breeding [66]. We are now starting to use also single nucleotide polymorphisms (SNPs) molecular markers that are more abundant in the genome and ame‐

Molecular markers are being used to achieve two main research objectives. First we are us‐ ing molecular markers to evaluate the progress obtained in conserving or increasing diversi‐ ty through participatory breeding, as already described in the section 2.3 of PPB success evaluation [42]. The genetic diversity of the newly introduced maize and beans landraces in‐ to the participatory plant breeding net is now also routinely characterized, with 20 to 22 SSR uniformly distributed throughout the maize and bean genomes respectively. This method enables us to check if sufficient diversity is present to allow selection and to select the most promising landraces in order to increase the genetic diversity by crossing genetically distant landraces. These studies have also allowed us to compare the genetic diversity with quality clustering of landraces [34]. In detail, 46 traditional maize landraces collected from known

should be taken into consideration on future participatory breeding selection.

**3.3. Development of molecular tools for assisting quality selection**

geographical identification of certain plant or crop product.

nable to automation for high-throughput genotyping [67].

still given its first steps.

274 Plant Breeding from Laboratories to Fields

Second, we are developing genetic studies to identify the genes responsible for our quality traits of interest and subsequently develop molecular markers that target those genes and that can be useful for marker assisted selection (MAS). Quality parameters for bread mak‐ ing, such as technological, nutritional and organoleptic traits, are generally characterized by a continuous variation. This continuous variation suggests the influence of several genes, and because of that, it is difficult to grasp by breeders and farmers. It is expected that sever‐ al of the maize bread quality parameters show quantitative inheritance. The identification of molecular markers that are linked to the controlling genes will be very helpful for the indi‐ rect selection through MAS of these complex quality traits. Marker-assisted selection is a powerful tool for the indirect selection of difficult traits at an early stage, before production of the next generation, thus speeding up the process of conventional plant breeding and fa‐ cilitating the improvement of traits that cannot be improved easily by conventional methods (reviewed by [67]). The identification and location of genes controlling quantitative traits through Quantitative Trait Loci (QTL) analysis has already been successful undertaken on maize nutritional quality [68, 69, 70].

In our genetic studies we started to use a marker-trait association analysis based on biparen‐ tal populations, where only a few target traits can be mapped within each population. It was possible by using this approach to identify several genomic regions responsible for ear fas‐ ciation related traits, widely present in the Portuguese maize landraces (on going research). Nevertheless, to be able to use this information for indirect selection of fasciated pheno‐ types, several runs of MAS and population development would be needed to narrow down the genomic regions. This method is time-consuming, but very powerful for the genes with large effect and the alleles with low frequency [71].

Another approach to identify molecular markers for using in MAS is association mapping based on linkage disequilibrium (LD). The availability of high-throughput genotyping tech‐ nology, together with advances in DNA sequencing and the development of statistical methodology appropriate for genome wide mapping analysis in the presence of considera‐ ble population structure, in species such as maize, contributed to an increased interest in LD association mapping [72]. This is the approach that we are now following for the quality ge‐ netic studies.

Unlike conventional biparental mapping populations, the natural populations used on this type of linkage analysis, are the products of many cycles of recombination and have the po‐ tential to show enhanced resolution of QTLs. Success depends on population size, control of population structure and the degree of LD in the population. LD levels vary both within and between species [73]. With this approach, marker–trait association is only expected when a QTL is tightly linked to the marker, because the accumulated recombination events occur‐ ring during the development of the lines will prevent the detection of any marker–trait loose association. In maize, the application of this approach has demonstrated the association be‐ tween several candidate genes and kernel composition traits, starch pasting properties and amylose levels [74].

Using SSR markers, the genetic diversity among inbred lines derived from the Portuguese germplasm collection was evaluated and compared with worldwide maize inbreds repre‐ sentatives [36]. The Portuguese inbred lines have maintained a level of genetic diversity sim‐ ilar to the foreign lines. Moreover, it was concluded that they are derivatives of miscellaneous populations, showing high genetic diversity and consequently representing a potential valuable source of interesting genes to introduce into modern cultivars [36], and a valuable germplasm for association studies.

Until now, no LD analysis or association studies were undertaken on the group of inbred lines of Portuguese origin, neither the identification of genes/QTLs controlling bread mak‐ ing ability. Presently, in order to address this gap, the collection of Portuguese maize inbred lines, derived mainly from Portuguese landraces, is being genotyped using microsatellites to detect population structure and to study LD.

Currently, the national efforts are focused on the study of the genetic control and the envi‐ ronmental effect on the antioxidant and aroma compounds as well as the bread making abil‐ ity. This study applies an association mapping approach using the previously characterized inbred lines that differ for endosperm types and colors. QTL associated candidate genes will be identify on the basis of positional information of the recently maize cloned genes (re‐ viewed by [75]). Candidate genes will be validated on the enduring landraces [7] and mod‐ ern improved OPVs (VASO project) that are also being characterized at genetic, nutritional, organoleptic (aroma volatiles) composition and antioxidants bioactivity. Specific molecular markers tightly linked to the identified QTLs will be identified or developed to provide breeders and farmers user-friendly markers to select for superior genotypes for quality maize bread. Additionally, it will allow the exploration of maize local resources and natural quality diversity in the reinvention of traditional maize to produce modern high quality bread with potential health benefits.

#### **3.4. Testing of higher quality experimental cultivars**

Nowadays, the most promising maize populations at agronomic, molecular and quality lev‐ el, collected during the 2005 expedition, are being evaluated and selected under a participa‐ tory approach in 13 different locations. This field research has been done in articulation with the original VASO project locations and improved populations, and now is under the super‐ vision of the ESAC researchers. The association of the farmers' perception with the newly available molecular and quality data can be extremely valuable to aggregate or separate populations, creating possible pools with heterosis that will be very useful to generate new populations or inbred lines. All the molecular and quality evaluations performed on these materials are being developed by researchers at ITQB/UNL and INIAV.

After a detailed characterization of the agronomic, genetic, nutritional, organoleptic and technological quality traits of 41 initial maize OPV (from the collecting expeditions plus VASO project), the most interesting materials were selected for the development of hybrid populations with specific quality traits and maintaining genetic diversity. Hybrid popula‐ tions can contribute to yield improvement and to avoid the collapse of some interesting germplasm. Dialel tests of the best materials are providing indications regarding heterosis among the chosen germplasm. These new populations are now under field evaluations. New synthetic populations with increased precocity are also being developed and are based upon the most superior Portuguese maize OPVs at agronomic level plus some American populations. These synthetic populations are also under field trial evaluation/selection at different farmers fields. Molecular and quality evaluations will follow on all these new de‐ veloped materials to sustain their improvement.
