**8. Pre-breeding and breeding perspectives to maize-AMF symbiosis**

One of the pivotal paths towards climate resilience and reversing the predicted negative impacts on food security is the adoption of climate-smart breeding approaches to design of high-yielding crops adapted to climate disturbances, such as increased abiotic and biotic stress [155–157]. Considering the tremendous positive impact of AMF symbiosis on maize yield, biotic and abiotic stress tolerance, Carbon and Nitrogen sink capacity, greenhouse gas mitigation [10, 14, 90, 91, 141, 158–161], it is crucial to improve the symbiotic response capacity in the crop as a contribution to food security in the face of the changing climate scenarios and continuous land degradation [117, 162]. Besides, beneficial microorganisms such as AMF are not just traits influencers but rather are part of the phenotypic expression of plants in an integrated system started from nutrient mobilization to resource

allocation to different plant functions, including resistance to biotic and abiotic tolerance, and nutrient efficiency and accumulation of assimilates in plant reserve organs which translates into yield [11, 163]. The current scientific information is more inclined towards a non-antagonistic impact of modern breeding activities on the mycorrhizal symbiotic capacity of maize [70, 96, 164–166], indicating the feasibility of efficient incorporating AMF-maize collaboration into ongoing breeding programs [165, 167, 168]. Therefore, the necessary continuous increase in crop yield to meet future food demands requires breeding efforts to improve symbiosis between critical crops such as maize with AMF [11, 166, 168].

For breeding to be efficient and achieve the expected genetic gains and progress, it is necessary to generate a host of useful pre-breeding information and devise breeding strategies based on state-of-art technologies selected through careful consideration of the factors that control maize-AMF interactions. In the coming sections, we review the critical pre-breeding information generated so far regarding maize symbiotic response to and interaction with AMF and identify areas necessitating further research to support breeding activities. We also devise breeding strategies to improve maize varieties' symbiotic response and interaction to AMF based on the current knowledge of maize-AMF interactions.

### **8.1 Genetic diversity and inheritance of maize response to AMF symbiosis**

The first and foremost step towards climate-smart crops is harnessing genetic diversity to allow the selection of superior material for breeding [169]. The maize genotype affects AMF abundance in the soil [170]. Maize has high genetic variance in response to AMF colonization [11, 70, 164, 170], indicating the possibility of selection in breeding. However, little information is available regarding the type of gene action controlling the maize response to AMP symbiosis, hence the need for more research in this area to judge opportunities for trait improvement through breeding. The high genetic diversity revealed by the few studies that investigated this pre-breeding characteristic of the maize-AMF interaction should be verified in several other genetic and environmental backgrounds to confirm promises of fast breeding progress. However, breeding progress might be hampered by the generally low to moderate heritability of maize responsiveness to AMF colonization [171].

Expectations of slow breeding progress are especially true for the traditional phenotypic selection, which strongly relies on phenotyping, increasing cost and time for budget-constrained breeding programs. Also, information related to the genetic control of maize-AMF interactions as to the proportion of additive and non-additive gene action involved in maize symbiotic response to AMF colonization [11]. Understanding the genetic control of any trait is crucial to designing effective breeding strategies for its improvement [172]. Therefore, besides the need to conduct more studies targeting maize levels of genetic diversity in response to AMF colonization, determining the preponderance of additive vs. dominance or epistasis genetic control on the trait needs to be elucidated better to inform future breeding strategies. New breeding approaches, especially those relying on advances in marker technologies such as marker-assisted selection (MAS) and genomic selection (GS), but also transgenic and genome editing (GE) techniques, could help to accelerate the improvement of maize responsiveness to AMF colonization and increase yield and stress tolerance [173–176].

### **8.2 Genetic architecture of maize response to AMF**

One of the first steps into implementing molecular breeding for maize response to AMF to increase yield performance is identifying genetic polymorphisms

*Climate-Smart Maize Breeding: The Potential of Arbuscular Mycorrhizal Symbiosis… DOI: http://dx.doi.org/10.5772/intechopen.100626*

controlling the final maize benefit from the symbiosis [177]. However, very few studies undertook to map genomic regions associated with maize response to AMF symbiosis [168, 171]. Kaeppler et al. [164] conducted the first quantitative trait loci (QTL) mapping for maize interaction with AMF in a population generated from a cross between B73 and Mo17. They identified one QTL controlling maize responsiveness to AMF, and such a low number of QTL was attributable to the low heritability of the trait in their study. Twenty years later, Ramírez-Flores et al. [161] undertook another study to identify QTL that determined maize benefits from AMF symbiosis. Several QTL were identified in this study, suggesting a polygenic nature in the control of the trait, contrary to the monogenetic direction indicated by the earlier study [168, 171]. Considering the molecular complexity involved in the symbiosis process, from the recognition between AMF and plant to the effective establishment of the symbiotic relationship [158], the polygenic nature of Maize-AMF interaction is more plausible. However, more studies are required to confirm this hypothesis further.

Confirming the genetic architecture of maize-AMF interaction is pivotal since effective conventional or molecular breeding strategy design will depend on the number, siege of QTL controlling the traits and their interactions [178, 179]. These studies should be conducted in a wide variety of germplasm and geographical backgrounds to discover a comprehensive number of QTL that could accurately determine the genetic architecture of the trait through meta-analyses and other integrative studies [180].

### **8.3 Research perspectives and breeding strategies for improved maize response to AMS**

Plant breeders generally are biased towards direct phenotypes, ignoring that most of these traits are mediated by beneficial microorganisms [11, 163, 167]. Although evidence points more towards a positive co-existence between modern plant breeding activities and practices, it is necessary to ascertain this status on target environments and maize populations. It is evident that response to AMF colonization is dependent on available resources such as soil phosphorus, crop species, and genotype [181]. The quantity and quality of soil phosphorus available to a particular crop are parameters that determine the maintenance of the diversity and quantity of the AMF community and their symbiotic capacity with maize [182]. Reports exist about the possibility of inhibition of the symbiosis between maize and AMF after artificial fertilization through the addition of external Phosphorus [183], making it necessary to adapt crop improvement for symbiotic capacity to target environments and cropping systems. **Figure 1** shows the cascade of pre-breeding and breeding activities that should be involved in a strategic crop improvement program targeting improved maize response to AMF colonization and symbiosis.

A typical breeding program for any trait should identify adequate germplasm, possibly including wild relatives, exotic accessions, and landraces, as a base population for breeding through careful mating designs and accelerate genetic gains towards possible variety release [184, 185]. The base population should be both phenotyped for target traits and genotyped with molecular markers to allow measuring phenotype and marker-based genetic diversity and population structure to optimize downstream research and breeding activities such as parent selection and cross designs for pre-breeding activities such as inheritance and genetic control studies, QTL mapping, and selection techniques such as phenotypic selection (PS), MAS, GS [186–189]. Phenotypic selection is a group of breeding methods basing the selection of superior genotypes for the next generation of for variety release on their observed phenotypic values. Phenotypic selection is best for highly

### **Figure 1.**

heritable and easy-to-measure traits. Both MAS and GS are based on selecting genotypes using molecular markers, albeit they are essentially different. MAS relies on mapped QTLs for a particular trait, of which it uses associated markers to select desired phenotypically unobserved lines. MAS works best when the trait is monogenic or oligogenic (controlled by one or a few large-effect QTL) [178, 190]. GS uses whole-genome markers to compute genomic-estimated breeding values of phenotypically unobserved genotypes as a basis for selection. GS performs best on polygenic traits that are controlled by multiple small-effect QTL, which characterizes most traits that plant breeders investigate [191–193].

In the case of maize interaction with AMF, phenotyping should be done with a control (non-mycorrhized plants) experiment to allow direct estimation of benefits offered by AMF symbiosis on traits of interest as in several studies [109, 194–196]. A comprehensive number of phenotypic, biochemical, and omic traits should be selected for phenotyping based on their direct or indirect involvement in or them being influenced by maize-AMF interactions to run univariate and multivariate analyses for genotype ranking, estimation of AMS effect on target traits, and strength and direction of relationships among traits [197]. Where possible, highthroughput phenotyping (HTP) techniques should be used to precisely measure and allow the accurate estimation of genetic and genomic parameters, including genetic control and inheritance, marker-trait association, and genomic prediction accuracies [198–200]. During the last decade, HTP technologies served to precisely measure the shoot biomass of tomato, barley [201], and Medicago [201, 202] growth trends under AMF colonization and to estimate nitrogen use efficiency of tomato, barley, and Medicago plants [203].

It is noteworthy that the inherent low allele diversity and low recombination rates arising from the bi-parental nature of such mapping populations and the short timespan from their generation to advanced stages used for mapping are critical

*Pre-breeding and breeding pipeline for improved maize-AMF symbiosis.*

### *Climate-Smart Maize Breeding: The Potential of Arbuscular Mycorrhizal Symbiosis… DOI: http://dx.doi.org/10.5772/intechopen.100626*

limits to most of these studies based on genetic linkage based QTL mapping methods [204–207]. The low statistical power is because all the genetic and allele diversity only comes from the two parents crossed to generate the mapping population. The low resolution of QTL is caused by the short time for creating such populations, which, even with recombinant inbred lines, is still too little to allow enough recombination in the genome of the lines [206, 207]. These limitations lead to low statistical power for QTL discovery and low resolution of the genomic regions mapped [206, 207]. These shortcomings could have partly explained the low QTL number mapped by Kaeppler et al. [164], and that results from Ramírez-Flores et al. [161] might not have comprehensively captured the genetic architecture of maize-AF interaction. Genome-wide association studies (GWAS) is an alternative and complementary technique to pipe rental population-based QTL mapping from which it differs by the reliance on populations composed of diverse lines with historical recombination events. Consideration of genome-wide association studies (GWAS) should allow complementing traditional bi-parental QTL analyses, especially in Joint Linkage Association Mapping (JLAM), a technique that combines the strengths of both GWAS and biparental QTL mapping to alleviate their respective weaknesses [205]. Also, GWAS will increase the statistical power and resolution of the resulting QTL [204, 205, 208].

One of the main challenges breeders face is combining several traits of interest in elite lines due to the pervasive pleiotropic effect and close-linkage of genes controlling these traits, two genetic phenomena that yield similar phenotypic outcomes but are difficult to distinguish between each other unless specific analyses are performed [185, 209]. A prerequisite for efficiently achieving multiple-trait selection is delineating the genetic basis of the correlations among traits through multivariate analyses [210]. Several multivariate GWAS and GS exist in the perspective of genomics-aided multi-trait selection for maize response to AMF symbiosis. Multivariate methods allow leveraging shared genetic information among traits and possibly environments to increase statistical power and accuracy [210, 211]. Also, GWAS could complement GS by including GWAS-discovered QTL as fixed effects in GS models, which is reported to improve prediction accuracy, thereby increasing genetic gains per unit time [173].

Since its invention, GWAS has evolved, moving from single-locus single-trait mixed linear models proposed by Yu et al. [212] to multi-locus multi-traits algorithms, which, unlike the former, jointly test associations between several traits and all genome-wide markers [213, 214]. Single-trait mixed linear models suffer from several weaknesses, including high rates of false-negative associations caused by multiple testing issues that require stringent Bonferroni thresholds [215]. In contrast, multi-locus multi-traits algorithms have better statistical power by avoiding correcting for multiple testing [214, 216]. However, these methods are still inefficient in differentiating between the two causes of trait correlations [216]; instead, integration of structural equation modeling (SEM) to GWAS is necessary [217]. Several GWAS packages that incorporate SEM are available for use in the case of maize-AMF interactions, for instance, GW-SEM [218], SEM-GWAS [219], GenomicSEM [220]. A more advanced software package is the multi-trait multilocus Structural Equation Modeling (mtmlSEM) that considers, besides the multitrait framework, a multi-locus approach to model associations between multiple traits and all loci simultaneously using SEM [221]. Also, GWAS results should be complemented with a robust candidate gene discovery and In Silico and lad-based prioritization steps to allow selection of high-confidence trait-associated genes that could be used in molecular breeding techniques such as MAS GE [222–225]. GE is a novel molecular breeding technique that, after mapping a genome region with an unfavorable genetic effect or with the potential of improving a trait, is used to precisely modify, insert, replace, or delete DNA in a genome [226].

Determining the genetic architecture of maize-AMF interactions will allow breeders to decide what selection approach would yield better genetic gains in a shorter time with a competitive budget requirement. However, considering the complexity of the molecular basis of symbiosis (see Section 7 of this chapter) and the probable polygenic nature of the phenomenon [168], it is expected that PS or MAS might not be efficient [227–230]. GS, especially combined with HTP technologies, should accelerate genetic gains while reducing overall variety development costs [231]. For complex and polygenic traits such as maize-AMF interactions subject to several non-genetic influences, multi-trait GS models, especially those considering multi-environment trials (MET) such as R packages BMTME [232], would be of tremendous benefit. Multi-trait multi-environment GS methods are being routinely used for diverse traits of diverse crops, including maize [233–236].
