Next Article in Journal
Simulation of Photosynthetic Quantum Efficiency and Energy Distribution Analysis Reveals Differential Drought Response Strategies in Two (Drought-Resistant and -Susceptible) Sugarcane Cultivars
Next Article in Special Issue
White Lupin Adaptation to Moderately Calcareous Soils: Phenotypic Variation and Genome-Enabled Prediction
Previous Article in Journal
What Makes the Life of Stressed Plants a Little Easier? Defense Mechanisms against Adverse Conditions
Previous Article in Special Issue
Identification of Candidate Genes for Rind Color and Bloom Formation in Watermelon Fruits Based on a Quantitative Trait Locus-Seq
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Plants
Volume 12
Issue 5
10.3390/plants12051041
plants-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Genome-Wide Association Study and Genomic Prediction for Fiber and Sucrose Contents in a Mapping Population of LCP 85-384 Sugarcane

by
Haizheng Xiong
1,†,
Yilin Chen
1,†,
Yong-Bao Pan
2,* and
Ainong Shi
1,*
1
Department of Horticulture, University of Arkansas, Fayetteville, AR 72701, USA
2
USDA-ARS, Sugarcane Research Unit, Houma, LA 70360, USA
*
Authors to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Plants 2023, 12(5), 1041; https://doi.org/10.3390/plants12051041
Submission received: 23 January 2023 / Revised: 11 February 2023 / Accepted: 21 February 2023 / Published: 24 February 2023
(This article belongs to the Special Issue Molecular Markers and Molecular Breeding in Horticultural Plants)
Browse Figures
Versions Notes

Abstract

:
Sugarcane (Saccharum spp. hybrids) is an economically important crop for both sugar and biofuel industries. Fiber and sucrose contents are the two most critical quantitative traits in sugarcane breeding that require multiple-year and multiple-location evaluations. Marker-assisted selection (MAS) could significantly reduce the time and cost of developing new sugarcane varieties. The objectives of this study were to conduct a genome-wide association study (GWAS) to identify DNA markers associated with fiber and sucrose contents and to perform genomic prediction (GP) for the two traits. Fiber and sucrose data were collected from 237 self-pollinated progenies of LCP 85-384, the most popular Louisiana sugarcane cultivar from 1999 to 2007. The GWAS was performed using 1310 polymorphic DNA marker alleles with three models of TASSEL 5, single marker regression (SMR), general linear model (GLM) and mixed linear model (MLM), and the fixed and random model circulating probability unification (FarmCPU) of R package. The results showed that 13 and 9 markers were associated with fiber and sucrose contents, respectively. The GP was performed by cross-prediction with five models, ridge regression best linear unbiased prediction (rrBLUP), Bayesian ridge regression (BRR), Bayesian A (BA), Bayesian B (BB) and Bayesian least absolute shrinkage and selection operator (BL). The accuracy of GP varied from 55.8% to 58.9% for fiber content and 54.6% to 57.2% for sucrose content. Upon validation, these markers can be applied in MAS and genomic selection (GS) to select superior sugarcane with good fiber and high sucrose contents.

1. Introduction

Sugarcane is a commercially important crop in tropical and subtropical regions of the world, accounting for more than 80% of global sugar production [1]. In 2021, the global sugarcane plantation area was 27.49 million hectares, and sugarcane production reached 1988.3 million tons [2]. As a renewable resource, sugarcane has been the best energy crop, available as food (sugar, Jaggery, syrup), feed (green tops/leaves) and fertilizer (pressed mud) [3]. In recent years, researchers have analyzed the potential of residues from the processing industry (bagasse, leaves and shoot-tips) for the production of second-generation ethanol and cogeneration of electric power. The production of alternative energy sources and the establishment of the concept of bio-refining have also led to a rapid increase in the global demand for sugar cane [4]. To meet this growing demand, the development of new cultivars with high biomass and sugar yields is essential. Modern sugarcane varieties (Saccharum spp. hybrids) are an interspecific hybrid of Saccharum officinarum (2n = 80) and S. spontaneum (2n = 40–128) [5]. S. officinarum has a high sugar content, while S. spontaneum provides resistance to various diseases and abiotic stresses [6,7].
The process of improving sugar content in interspecific sugarcane hybrids through backcrossing, known as “Nobilization”, requires 3 to 6 generations of crosses to stabilize the genome and restore the high sugar trait [8]. This process initiated with a (2n + n) chromosome transmission mode and ends with =100–144 chromosomes [9]. Genome analysis via in situ hybridization shows that S. officinarum and S. spontaneum make up about 80% and 10–20% of the genome of modern cultivars, respectively, with 10% being recombinant chromosomes [7,9,10]. The total genome size of sugarcane hybrids is estimated to be about 10 Gb with 10 homologous linkage groups [11]. The complex polyploid and aneuploid feature of sugarcane hybrids presents challenges in genetic analysis and trait isolation, which can be addressed through the use of advanced statistical methods and genomic tools in breeding [12]. Current breeding efforts focus on improving sugar yield, disease and pest resistance, ratooning ability, cold tolerance and total biomass yield [3]. However, the development of a new sugarcane cultivar typically takes 12–15 years and involves annual evaluation of 60,000 to 250,000 seedlings [13,14]. The meiotic chromosome pairing in sugarcane is primarily bivalent, but within specific homologous groups, there is a complex and unbalanced pattern of mutual and systematic pairing [15]. Using molecular markers associated with relevant agronomic traits can aid in the selection of suitable parent plants and speed up the process of genetic improvement during breeding, ultimately reducing the time and cost of developing new varieties [16].
Linkage/family-based genetic mapping has been successful in identifying several quantitative trait loci (QTL) that have major or minor phenotypic effects on for both simple and complex traits [17]. However, linkage mapping has some limitations, including the difficulty in establishing a single population that simultaneously isolates multiple traits, the time-consuming process of establishing mapping populations, the lack of high-density mapping, and the overestimation of phenotypic variations by markers [15]. Despite the numerous reports on linkage analysis during the past three decades, few QTLs have been identified in sugarcane due to the lack of a clear allelic isolation pattern in sugarcane genomes, such as complete polysomy or complete disomy. Family-based linkage mapping mainly utilizes F1 populations, which can be limited in clarity due to the large number of genes, polyallelic and/or polygenic properties, underutilization of sugarcane gene pools and limited internal comparisons between genes [15,18]. These limitations can be overcome through marker-trait association (MTA) studies based on linkage disequilibrium (LD) using diverse genotypes that capture a wide range of allelic diversity, often referred to as association mapping (AM) [16,19]. Linkage disequilibrium mapping was initially used in human genetic studies as an alternative to marker-trait association identification in plants [20,21]. This approach has been found to be particularly useful in sugarcane due to the large number of LD present in the genome [22]. Researchers have used models corrected for population structure (Q) to identify markers for disease resistance and other traits [23]. For example, Wei et al. (2006) identified markers for resistance to smut, Pachymetra root rot, scalding and Fiji leaf gall using Q-modified models [24]. Similarly, using a Q-modified model, the significant MTA of sugarcane yield and sugar content were 43% and 38%, respectively [16]. Debibakas et al. (2014) identified six independent markers of sugarcane yellow leaf virus resistance using a mixed linear model (MLM) of Q-K (population structure—kin) combination [25]. Gouy et al. (2015) used the Q-K combination to evaluate the general linear model (GLM) and MLMs of MTA for 13 traits related to agricultural morphology, sucrose yield, bagasse content and disease resistance [23]. Banerjee et al. (2015) evaluated the role of MTA on various sucrose and yield traits [26]. In another study, Q-K analysis was used to identify four SSR markers associated with red rot resistance [27]. Recently, Ukoskit et al. (2019) used MLM on a diversity panel consisting of 200 germplasm accessions and identified two SSR markers associated with polarization (Pol) and sugar production [28]. SNP (Single Nucleotide Polymorphism) markers have been used and have shown great potential in sugarcane studies in sugarcane studies of genetic basis of various traits. The use of SNP markers has allowed researchers to identify specific DNA regions that are associated with these traits, providing insights into the genetic mechanisms underlying these important characteristics with higher genomic coverage [29]. However, sampling bias, technical requirements and the cost of SNP genotyping still represent barriers for some sugarcane researchers, particularly in low-resource settings [30,31,32].
Genomic prediction (GP) is a method of predicting the breeding value of an individual plant based on its genomic data [17,33]. GP has the potential to significantly expedite the breeding process by allowing breeders to select for traits of interest without the need for extensive phenotyping and field testing [34]. In a study by Olatoye and colleagues, the effectiveness of GP and MAS in predicting traits with different genetic structures and marker densities was evaluated. Both F1 and BC1 populations were used for lignocellulosic, biomass and disease resistance analyses. The results showed that GP had a higher prediction accuracy and the best performance in simulating character values. GP was able to identify more genotypes, with prediction accuracy for characters reaching 44–77% [35]. Deomano and colleagues conducted a selection experiment in sugarcane using three different commercial populations at different stages. They found that the genomic prediction model with marker data provided a higher prediction accuracy (25–45%) than the model with pedigree data alone [36]. Similarly, in a study involving 3984 individuals, Hayes et al. showed that GP had a prediction accuracy of 30–47% for cane yield, commercial cane sugar (%), fiber content and flowering traits [37]. Islam and colleagues used target enrichment sequencing to generate 8825 SNP markers from 432 sugarcane clones and found that the prediction accuracy of various GP models for brown and orange rust resistance ranged from 0.28 to 0.43 and 0.13 to 0.29, respectively. The inclusion of a known master gene for brown rust resistance as a fixed effect in the GP model also significantly reduced the minimum number of markers and the size of training populations [38].
A basic premise of genomics-assisted breeding is the identification of trait-associated molecular markers. A genetic linkage framework map containing AFLP, TRAP and SSR markers has been constructed using a self-progeny mapping population of LCP 85-384, the most popular sugarcane cultivar in Louisiana from 1997 to 2007 [39]. The objectives of this study were to perform genome-wide association to identify DNA markers associated with sucrose and fiber contents in the LCP 85-384 mapping population, and to assess the efficiency of genomic prediction in order to apply marker-assisted genomic selection in sugarcane breeding programs.

2. Results

2.1. Fiber and Sucrose Contents

The sucrose and fiber contents of the 237 clones were phenotyped for two years (Table S1). The statistical data and the distribution of the phenotypic traits are presented in Table 1 and Figure 1, respectively. The average of these values in 2006 was consistently higher than those in 2007, while the variance, standard error and coefficient of variation for the two traits were larger in 2007; however, these data were still at a low level in both years. The estimated broad-sense heritability values (h2) were high in both years (0.65~0.67 for sucrose and 0.74~0.77 for fiber). The same conclusion can be drawn from the population distribution of the two phenotypic data collected over two years in Figure 1. According to correlation analysis, sucrose and fiber contents were moderately negatively correlated (−0.23 and −0.40) for both years.

2.2. Genetic Relationship and Population Structure

The population structure of the 237 self progenies was initially inferred using STRUCTURE 2.3.1 and the peak of delta K was observed at K = 3, by 135 polymorphic markers (Table S2) indicating the presence of three sub-populations (Figure 2A, Table S1). At a threshold value of 0.5, 41 of the 237 self progenies (17.2%) were assigned to the Q1 subpopulation; 47 self progenies (19.7%) were assigned to Q2; 39 self progenies (16.4%) assigned to Q3; Q1Q2 and Q2Q3 (Qx + Qy > 70%) were both assigned by 11 self progenies (4.6%); 26 self progenies (10.9%) were assigned to Q1Q3; and 62 self progenies (26.1%) were assigned to QX (Qx ≈ Qy ≈ Qz) (Figure 2B). Phylogenetic analysis and a population admixture map of the 237 self progenies using MEGA 11 software also showed a clustering pattern consistent with that inferred by structure K = 3 (Figure 3A). The most closely related self progenies based on Structure analysis were grouped in the neighbor branches of the phylogenetic tree by MEGA analysis. The three groups were also observed based on PCA dimensions (Figure 3B). Therefore, the 237 self progenies can be divided into three sub-populations based on both structural and phylogenetic analysis.

2.3. GWAS Analysis

GWAS analysis was conducted by using 1,310 polymorphic alleles (260 SSR, 950 AFLP and 100 TRAP) with four models: three models (SMR, GLM, MLM) in TASSEL 5 and FarmCPU in GAPIT 3. The QQ plots showed a large divergence from the expected distribution, indicating that there were alleles associated with both sucrose and fiber contents in the LCP 85-384 mapping population (Figure 4). In this study, 25, 18 and 25 alleles were significantly associated with fiber content (LOD > 2.0) of 2006, 2007 and the mean of 2006 and 2007, respectively; 13, 21 and 14 alleles were significantly associated with sucrose content of 2006, 2007 and the mean of 2006 and 2007, respectively, by using GLM analysis in TASSEL 5 (Table S3). Using MLM, 10, 11 and 11 alleles were associated significantly with fiber content of 2006, 2007 and mean of 2006 and 2007, respectively; and 7, 10 and 6 alleles were significantly associated with sucrose content of 2006, 2007 and the mean of 2006 and 2007, respectively (Table S3). As in SMR, 12, 15 and 12 markers were significantly associated with sucrose content of 2006, 2007 and mean of 2006 and 2007, respectively; 31, 19 and 23 markers were significantly associated with fiber content of 2006, 2007 and the mean of 2006 and 2007, respectively (Table S3). In addition, 8, 8 and 11 markers were significantly associated with sucrose content of 2006, 2007 and the mean of 2006 and 2007, respectively, and 5, 4 and 10 markers were significantly associated with fiber content of 2006, 2007 and mean of 2006 and 2007, respectively, using FarmCPU analysis (Table S3). After comprehensively measuring the LOD values and repeatability of the associated alleles across different models and years, 9 and 13 associated alleles were selected as association markers for sucrose and fiber contents, respectively (Table 2), which can be used in molecular breeding.

2.4. Genomic Prediction Analysis

In this study, two marker datasets, namely, the All-allele dataset and the Trait-associated-allele dataset, were used for GP analysis by five models: BA, BB, BL, BRR and rrBLUP (Figure 5). The predictive ability of each marker dataset was estimated using the sucrose and fiber content data in 2006, 2007 and the 2006 and 2007 mean. For sucrose contents, the All-allele dataset produced similar average accuracies among all models ranging from 16.4% (BB) to 20.1% (BA), as well as the average accuracies of fiber content ranging from 17.9% (BB) to 25.3% (BA). The average accuracies of the GWAS-associated allele set ranged from 55.8% (BB) to 58.9% (BA) for fiber contents and 54.6% (BB) to 57.2% (BA) for sucrose contents (Table S4). Compared with the All-allele dataset, the Trait-associated-allele set had higher accuracies in all models, 36.1% higher for fiber content and 37.4% higher for sucrose content, respectively. Therefore, using trait-associated marker alleles to perform GS is more efficient in selecting sucrose and fiber content in sugarcane breeding.

3. Discussion

3.1. Phenotyping

The cane stalks consist of a core containing most of the extractable sucrose, and the major component of outer layer is fiber containing lignocellulose. Improving sucrose or fiber content in cane stalks and balancing these two traits for practical industry have always been an important and challenging task for sugarcane breeders [40]. Increasing sucrose content while maintaining an acceptable level of fiber content in the stalk rather than increasing cane yield is seen as an economically viable option, as it can potentially avoid the increased costs associated with increased harvesting, transport and milling [1,41]. The heritability is crucial to predict the potential for improving those quantitative traits through selective breeding or other genetic interventions [42,43]. In this study, the broad-sense heritability estimates for the sucrose and fiber contents were high at 65~67% for sucrose and 74~77% for fiber, showing that genetic variation was a significant factor in the two traits, and indicating that it would be effective to improve the traits in selection. As expected, negative correlations (−0.23 and −0.40) were observed between sucrose and fiber contents during the two-year phenotyping experiments, indicating that the synthesis of sucrose in plants requires the breakdown of complex carbohydrates such as fiber [44].

3.2. The Self-Progeny Population of LCP 85-384

LCP 85-384 was a very popular cultivar in the Louisiana sugar industry due to its high sugar yield and various favorable agronomic traits, including tolerance to biotic and abiotic stresses. The sucrose and fiber contents of some self progenies were found to be higher than those of commercial sugarcane hybrids. The higher sucrose and fiber contents of LCP 85-384 may be attributed to the presence of S. officinarum, S. spontaneum and S. barberi clones in its pedigree [39]. LCP 85-384 has also been used extensively as a parent in Louisiana breeding programs. The genetic analysis of its self-progeny mapping population with AFLP, SSR and TRAP DNA markers has resulted in an enriched linkage map, and several alleles from these markers were found to be associated with sucrose or fiber content [45,46].

3.3. Genetic Diversity, Population Structure and PCA

Sugarcane is a complex polyploid plant species with a large genome size and a high level of genetic diversity [11]. The presence of sub-populations in a mapping population can create challenges for association studies, and methods such as STRUCTURE and PCA are often used to deal with false positives related to population structure [21]. These methods can be used to identify and control population structure in order to increase the power to detect true associations in a GWAS. It is also noted that the use of STRUCTURE may be limited in the case of sugarcane due to its complex polyploid genome and the lack of clear discontinuities in the population [16]. In this study, both PCA and STRUCTURE were used as a random component in the GWAS analysis to control for population structure and maximize the power to detect true associations. The results of the population structure analysis, the phylogenetic analysis and the distribution of genotypes on the PCA were all consistent with one another, suggesting a high level of consistency in the data.

3.4. Genome-Wide Association Study

The long breeding cycles in sugarcane have resulted in large blocks of linkage disequilibrium (LD), which can be advantageous for identifying markers associated with genes that regulate important agronomic traits [47]. However, the highly heterozygous genome of sugarcane and severe inbreeding depression for certain traits can make it difficult to study the F1 progeny of two heterozygous parents, or even the selfed progeny population of a selected cultivar [22]. That is why linkage mapping has limitations, such as inflating estimates and hard-to-find minor QTLs. Multiple models were developed, including SMR, GLM, MLM and FarmCPU, etc., for GWAS based on linkage disequilibrium. Nonetheless, the performance of these models can vary depending on the specific population, the trait being studied and the environment in which the study was conducted [48]. Previous research has shown that the differences in the performance of these models are often due to the interactions between the methods and other factors [49,50]. Each of these models has its own strengths and limitations, and researchers may comprehensively consider more models based on the specific research question and the data available. In this case, we applied four GWAS models to identify QTLs of fiber and sucrose contents with the LOD threshold > 2.0. These trait-associated QTLs (Table S3) were highly consistent with former reports, at 24/28 [45] in sucrose content and 2/3 in fiber contents [46]. Then, considering the consistency among models, 13 and 9 markers were selected as putative QTLs for fiber and sucrose contents, respectively. Of these, most of the alleles were AFLPs, while only two SSRs and one TRAP were associated with sucrose content and one SSR was associated with fiber. This is likely to be caused by differences in the number of markers and the level of polymorphism; therefore, the type of marker used can also affect the accuracy and reliability of GWAS results.

3.5. Genomic Prediction

Normally, genomic prediction (GP) does not work well for traits that are influenced by complex interactions between multiple genes and the environment. Therefore, the traits with high heritability, such as the sucrose and fiber contents in this study, would be more suitable for breeding value prediction by GP [51,52]. We selected the most popular and reliable GP models that have been applied to sugarcane, including one BLUP model and four Bayesian models in this case. The accuracies of GP based on a Trait-associated-allele set (Table S4) were higher than the All-allele set, which demonstrated the importance of significant alleles. It is worth noting that in all GP models, predictions for fiber content were higher than those for sucrose content, which is consistent with their heritability, demonstrating the reliability of the prediction.
The typical breeding cycle for sugarcane can vary depending on the specific cultivar, growing conditions and breeding method, but it can often last 12–15 years or even longer for a new cultivar, including time for seed production, growth, evaluation and selection [13,14]. In the last two years, researchers began to introduce the GP technology into sugarcane breeding, which involves many traits such as yield, bagasse, sucrose and disease resistance based on SNP markers [35,36,37,38]. For these applications, using DNA markers associated with specific traits of interest, breeders can identify individuals with desirable traits early in the growth cycle, before they become fully mature. This allows breeders to make their selections more quickly and helps reduce the time and labor consumption required to evaluate the crop. Additionally, GP and marker-assisted selection can help breeders to track progress more accurately while pursuing their breeding goals, as well as avoiding selections that are likely to negatively affect other traits. However, the use of large-scale sequencing or diversity array technology for molecular data analysis can be expensive and requires advanced technology. This study was the first to use SSR, AFLP and TRAP for GP analysis in sugarcane, which makes it more practical and reliable for breeding teams with limited resources [53,54].

4. Materials and Methods

4.1. Plant Materials

A mapping population of LCP 85-384 sugarcane with 237 genetically verified self progenies was used in this study. LCP 85-384 was developed from a cross between CP 77-310 and CP 77-407 through a collaboration among the Louisiana State University Ag Center, the USDA-ARS Sugarcane Research Unit, and the American Sugarcane League of the USA, Inc. [39].

4.2. Phenotyping and Analysis

Replicated field trials of the LCP 85-384 mapping population were conducted in 2006 (plant cane crop) and 2007 (first ratoon crop) at the Ardoyne Farm, USDA-ARS Sugarcane Research Unit, Schriever, Louisiana, USA (29°37′55.93″ N, 90°50′10.47″ W). In the field, each progeny was replicated and planted in a single 3 m-long plot, with 1.0 m between plots and 1.8 m between rows. Sucrose content and fiber data were measured by hand-cutting 10 millable stalks at ground level, removing the top just below the apical meristem, knife-stripping the stalks to remove leaf and sheath tissue, bundling and tagging the stalks and then weighing them to obtain an estimate of individual stalk weight (kg). The stalks were cut and shredded in a pre-breaker (Cameco Industries Inc., Thibodaux, LA, USA). Juice was extracted from a 1 kg shredded sample in a core press (Cameco Industries Inc., Thibodaux, LA, USA) under 211 kg/cm pressure in the juice quality laboratory to measure total soluble solids, Brix and sucrose content. The remaining “cake” of fibrous residues was weighed, then dried at 66 °C for 72 h to obtain the dry weight of the fibrous residues. The sucrose and fiber content was estimated according to the methods of Legendre and Henderson [55].
The phenotypic data were analyzed by SAS 9.2 (SAS 2008) software. The variance of fiber and sucrose contents was analyzed using the model of Aitken et al. (2006) [56]. Covariance analysis of sucrose and fiber contents was performed using various covariance components and mean product. The sources of variability between phenotypic datasets, as well as the system and error variances, as a proportion of the variance, were also calculated by dividing the system variance by the total variance for a particular trait. The formula of heritability (h2) was used to determine each trait.
h 2 = σ G 2 σ G 2 + σ ε 2 / r
where σ G 2 is the genetic variance, σ ε 2 the residual variance and r the number of replicates.

4.3. Genetic Markers and Genotyping

The details of DNA extraction, genotyping and linkage map construction were previously described by Andru et al. (2011) [39] and Liu et al. (2016) [45]. In brief, genomic DNA was extracted from young leaves from each progeny by Pan et al. (2000) [57]. DNA concentrations were measured at 260 nm using a NanoDrop 1000 spectrophotometer (NanoDrop, Bethesda, MD, USA) and then equilibrated by agarose gel electrophoresis. Genotyping protocols including primer sequence, PCR amplification and PCR product detection for AFLP and TRAP markers were described in Andru et al. (2011) [39] and for SSR genotyping in Pan et al. (2007) [58]. The construction of an enriched genetic linkage map of LCP85-384 was described in Liu et al. (2016) [45]. In this study, we picked up the markers with less than 5% missing data and more than 2% minor frequency as genotyping data. In total, 60 SSR, 63 AFLP and 12 TRAP markers were used, including 260 polymorphic SSR alleles (4.3/marker), 950 polymorphic AFLP alleles (15/marker) and 100 TRAP polymorphic alleles (8.3/marker) (Table S2).

4.4. Genetic Diversity and Population Structure

All the polymorphic SSR, TRAP and AFLP alleles scored on the 237 self progenies were used to estimate their genetic relationship. STRUCTURE 2.3.1 was used to infer the population structure [59]. To identify the number of populations (K) and capture the major structure in the data, we set up at a burn-in period of 10,000 Markov Chain Monte Carlo iterations and 100,000 run length with an admixture model following Hardy–Weinberg equilibrium, correlated allele frequencies, as well as independent loci for each run. Ten independent runs were performed for each simulated K value, ranging from 1 to 10. For each simulated K, the statistical value delta K was calculated using the formula described by Evanno et al. [60]. The optimal K was determined using Structure Harvester [61]. Genetic dissimilarities between all pairwise combinations of progenies were calculated using the Dice index described by Shi et al. [62]. Then, a Neighbor Joining tree was built from the matrix of pairwise dissimilarities using the software MEGA 11 [63]. Phylogenetic relationships and principal component analysis (PCA) were generated by TASSEL 5.2.13 to analyze genetic relationships among progenies and to determine the optimal number of clusters in the study. The number of principal components (PC) was chosen according to the optimum subpopulation determined in STRUCTURE 2, and a PCA plot was drawn using R package ggplot2 using the data from TASSEL 5 [64]. The genetic diversity was also assessed, and phylogenetic trees were drawn using MEGA 11 based on the Maximum Likelihood tree method with the parameters described by Shi et al. [62]. During the drawing of the phylogeny trees, the population structure and the cluster information were imported for the combined analysis of genetic diversity.

4.5. Genome-Wide Association Study and Genomic Prediction

GWAS was performed for fiber and sucrose content in the self-progeny mapping population of LCP 85-384 using single marker regression (SMR), a generalized linear model (GLM), the mixed linear model (MLM) [65] by Tassel 5 and the fixed and random model circulating probability unification (FarmCPU) [66] of R software GAPIT 3 [67]. The significant threshold of associations was set when the LOD [−log(p-value)] > 2.0 in this study.
GP was conducted using two types of genotype datasets: all polymorphic alleles (1310) and traits associated alleles according to the GWAS results. All the alleles with LOD > 2.0 (Table S3) were considered and applied for prediction with the phenotype of the corresponding year. Genomic estimated breeding value (GEBV) was computed using 5 different statistical models, namely, Ridge regression best linear unbiased predictor (rrBLUP) [68], Bayes ridge regression (BRR), Bayes A (BA), Bayes B (BB), and Bayesian least absolute shrinkage and selection operator (BL) [69]. A five-fold cross validation to a training/testing set as 80%/20% was performed for the genomic prediction study. The association panel was randomly divided into five disjointed groups. A total of 100 replications were conducted at each fold. Mean and standard errors corresponding to each fold were computed [70].

5. Conclusions

Fiber and sucrose contents and genotyping data of the 237 self progenies of an LCP 85-384 mapping population were analyzed in this study. The results showed that the two traits have large genetic variances with a high broad-sense heritability: 68–69% for sucrose content and 82–85% for fiber content. GWAS identified 9 and 13 marker alleles as associated with sucrose and fiber contents, and GP estimated up to 57.2% and 58.9% prediction accuracy for sucrose and fiber content, respectively, when using a trait-associated-allele dataset. The study identifies GWAS-derived marker alleles significantly associated with sucrose and fiber contents and provides useful information for the breeders to use in marker-assisted and genomic selection programs.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/plants12051041/s1, Table S1: Sucrose and fiber contents in two years (2006 and 2007) and structure analysis in K = 3, among 237 clones of LCP 85-384 population. Table S2: The profile of 135 markers including SSR, AFLP and TRAP with produced polymorphic fragments and coverage/allele. Table S3: List of all associated markers with sucrose and fiber contents by the data in 2006, 2007, and 2006 and 2007 mean in each of the four models, SMR, GLM, MLM and FarmCPU in LCP 85-384 population. Table S4: Genomic prediction (GP) mean accuracy (r-value) and Standard Error for fiber and sucrose using five GP models: Ridge regression best linear unbiased predictor (rrBLUP), Bayes ridge regression (BRR); ’Bayes A’ (BA), ‘Bayes B (BB)’ Bayesian least absolute shrinkage (BL) and selection operator based on All-allele set and Trait-associated allele set.

Author Contributions

Y.-B.P. conceived the project, designed the experiments, conducted the SSR fingerprinting and collected the original data. H.X., A.S. and Y.C. organized and analyzed the original data. H.X. drafted the manuscript. Y.-B.P. and A.S. critically revised the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by competitive grower/processor check-off grant funds administrated by the American Sugar Cane League of the U.S.A., Inc., Thibodaux, Louisiana, U.S.A. a USDA-ARS Non-Assistance Cooperative Agreement on Genetic Analysis and Trait-Specific Molecular Marker Development (Accession No. 440501); and the China Agriculture Research System of MOF and MARA (Grant No. CARS-17).

Data Availability Statement

The data presented in this study are available in this article and Supplementary Material.

Acknowledgments

We thank Lionel Lomax and Sheron Simpson for their excellent technical support. The authors are thankful to Perng-Kuang Chang (USDA-ARS, New Orleans, LA, USA) and Paul white (USDA-ARS, Houma, LA, USA) for reviewing the manuscript with excellent editorial comments. USDA is an equal opportunity provider and employer.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Waclawovsky, A.J.; Sato, P.M.; Lembke, C.G.; Moore, P.H.; Souza, G.M. Sugarcane for bioenergy production: An assessment of yield and regulation of sucrose content. Plant Biotechnol. J. 2010, 8, 263–276. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  2. FAOSTAT. Food and Agriculture Organization of the United Nations. 2021. Available online: https://www.fao.org/faostat/en/#data/QCL (accessed on 20 February 2023).
  3. Tew, T.L.; Cobill, R.M. Genetic improvement of sugarcane (Saccharum spp.) as an energy crop. In Genetic Improvement of Bioenergy Crops; Springer: New York, NY, USA, 2008; pp. 273–294. [Google Scholar]
  4. Landell, M.G.d.A.; Scarpari, M.S.; Xavier, M.A.; Anjos, I.A.d.; Baptista, A.S.; Aguiar, C.L.d.; Silva, D.N.d.; Bidóia, M.A.P.; Brancalião, S.R.; Bressiani, J.A. Residual biomass potential of commercial and pre-commercial sugarcane cultivars. Sci. Agric. 2013, 70, 299–304. [Google Scholar] [CrossRef] [Green Version]
  5. D’Hont, A.; Ison, D.; Alix, K.; Roux, C.; Glaszmann, J.C. Determination of basic chromosome numbers in the genus Saccharum by physical mapping of ribosomal RNA genes. Genome 1998, 41, 221–225. [Google Scholar] [CrossRef]
  6. Cordeiro, G.M.; Taylor, G.; Henry, R.J. Characterisation of microsatellite markers from sugarcane (Saccharum sp.), a highly polyploid species. Plant Sci. 2000, 155, 161–168. [Google Scholar] [CrossRef]
  7. Bremer, G. Problems in breeding and cytology of sugar cane. Euphytica 1961, 10, 59–78. [Google Scholar] [CrossRef]
  8. Ming, R.; Moore, P.H.; Wu, K.K.; D’Hont, A.; Glaszmann, J.C.; Tew, T.L.; Mirkov, T.E.; Da Silva, J.; Jifon, J.; Rai, M. Sugarcane improvement through breeding and biotechnology. Plant Breed. Rev. 2010, 27, 15–118. [Google Scholar]
  9. D’Hont, A.; Grivet, L.; Feldmann, P.; Glaszmann, J.; Rao, S.; Berding, N. Characterisation of the double genome structure of modern sugarcane cultivars (Saccharum spp.) by molecular cytogenetics. Mol. Gen. Genet. 1996, 250, 405–413. [Google Scholar] [CrossRef]
  10. Piperidis, N.; Piperidis, G.; D’Hont, A. Molecular cytogenetics. In Genetics, Genomics and Breeding of Sugarcane; CRC Press: Boca Raton, FL, USA, 2010; pp. 27–36. [Google Scholar]
  11. Grivet, L.; Arruda, P. Sugarcane genomics: Depicting the complex genome of an important tropical crop. Curr. Opin. Plant Biol. 2002, 5, 122–127. [Google Scholar] [CrossRef]
  12. Souza, G.M.; Berges, H.; Bocs, S.; Casu, R.; D’Hont, A.; Ferreira, J.E.; Henry, R.; Ming, R.; Potier, B.; Van Sluys, M.-A. The sugarcane genome challenge: Strategies for sequencing a highly complex genome. Trop. Plant Biol. 2011, 4, 145–156. [Google Scholar] [CrossRef]
  13. Cheavegatti-Gianotto, A.; de Abreu, H.M.C.; Arruda, P.; Bespalhok Filho, J.C.; Burnquist, W.L.; Creste, S.; di Ciero, L.; Ferro, J.A.; de Oliveira Figueira, A.V.; de Sousa Filgueiras, T. Sugarcane (Saccharum X officinarum): A reference study for the regulation of genetically modified cultivars in Brazil. Trop. Plant Biol. 2011, 4, 62–89. [Google Scholar] [CrossRef] [Green Version]
  14. Bischoff, K.P.; Gravois, K. The development of new sugarcane varieties at the LSU AgCenter. J. Am. Soc. Sugar Cane Technol. 2004, 24, 142–164. [Google Scholar]
  15. Banerjee, N.; Khan, M.S.; Swapna, M.; Singh, R.; Kumar, S. Progress and prospects of association mapping in sugarcane (Saccharum species hybrid), a complex polyploid crop. Sugar Tech 2020, 22, 939–953. [Google Scholar] [CrossRef]
  16. Racedo, J.; Gutiérrez, L.; Perera, M.F.; Ostengo, S.; Pardo, E.M.; Cuenya, M.I.; Welin, B.; Castagnaro, A.P. Genome-wide association mapping of quantitative traits in a breeding population of sugarcane. BMC Plant Biol. 2016, 16, 142. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  17. Wang, J.; Roe, B.; Macmil, S.; Yu, Q.; Murray, J.E.; Tang, H.; Chen, C.; Najar, F.; Wiley, G.; Bowers, J. Microcollinearity between autopolyploid sugarcane and diploid sorghum genomes. BMC Genom. 2010, 11, 261. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  18. Alwala, S.; Kimbeng, C.A.; Henry, R.; Kole, C. Molecular genetic linkage mapping in Saccharum: Strategies, resources and achievements. In Genetics, Genomics and Breeding of Sugarcane; CRC Press: Boca Raton, FL, USA, 2010; pp. 69–96. [Google Scholar]
  19. Barreto, F.Z.; Rosa, J.R.B.F.; Balsalobre, T.W.A.; Pastina, M.M.; Silva, R.R.; Hoffmann, H.P.; de Souza, A.P.; Garcia, A.A.F.; Carneiro, M.S. A genome-wide association study identified loci for yield component traits in sugarcane (Saccharum spp.). PLoS ONE 2019, 14, e0219843. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  20. Sorkheh, K.; Malysheva-Otto, L.V.; Wirthensohn, M.G.; Tarkesh-Esfahani, S.; Martínez-Gómez, P. Linkage disequilibrium, genetic association mapping and gene localization in crop plants. Genet. Mol. Biol. 2008, 31, 805–814. [Google Scholar] [CrossRef]
  21. Mackay, I.; Powell, W. Methods for linkage disequilibrium mapping in crops. Trends Plant Sci. 2007, 12, 57–63. [Google Scholar] [CrossRef]
  22. Raboin, L.-M.; Pauquet, J.; Butterfield, M.; D’Hont, A.; Glaszmann, J.-C. Analysis of genome-wide linkage disequilibrium in the highly polyploid sugarcane. Theor. Appl. Genet. 2008, 116, 701–714. [Google Scholar] [CrossRef]
  23. Gouy, M.; Rousselle, Y.; Thong Chane, A.; Anglade, A.; Royaert, S.; Nibouche, S.; Costet, L. Genome wide association mapping of agro-morphological and disease resistance traits in sugarcane. Euphytica 2015, 202, 269–284. [Google Scholar] [CrossRef]
  24. Wei, X.; Jackson, P.A.; McIntyre, C.L.; Aitken, K.S.; Croft, B. Associations between DNA markers and resistance to diseases in sugarcane and effects of population substructure. Theor. Appl. Genet. 2006, 114, 155–164. [Google Scholar] [CrossRef]
  25. Débibakas, S.; Rocher, S.; Garsmeur, O.; Toubi, L.; Roques, D.; D’Hont, A.; Hoarau, J.-Y.; Daugrois, J.-H. Prospecting sugarcane resistance to sugarcane yellow leaf virus by genome-wide association. Theor. Appl. Genet. 2014, 127, 1719–1732. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  26. Banerjee, N.; Siraree, A.; Yadav, S.; Kumar, S.; Singh, J.; Pandey, D.K.; Singh, R.K. Marker-trait association study for sucrose and yield contributing traits in sugarcane (Saccharum spp. hybrid). Euphytica 2015, 205, 185–201. [Google Scholar] [CrossRef]
  27. Singh, R.K.; Banerjee, N.; Khan, M.; Yadav, S.; Kumar, S.; Duttamajumder, S.; Lal, R.J.; Patel, J.D.; Guo, H.; Zhang, D. Identification of putative candidate genes for red rot resistance in sugarcane (Saccharum species hybrid) using LD-based association mapping. Mol. Genet. Genom. 2016, 291, 1363–1377. [Google Scholar] [CrossRef] [PubMed]
  28. Ukoskit, K.; Posudsavang, G.; Pongsiripat, N.; Chatwachirawong, P.; Klomsa-Ard, P.; Poomipant, P.; Tragoonrung, S. Detection and validation of EST-SSR markers associated with sugar-related traits in sugarcane using linkage and association mapping. Genomics 2019, 111, 1–9. [Google Scholar] [CrossRef]
  29. Gutierrez, A.F.; Hoy, J.W.; Kimbeng, C.A.; Baisakh, N. Identification of genomic regions controlling leaf scald resistance in sugarcane using a bi-parental mapping population and selective genotyping by sequencing. Front. Plant Sci. 2018, 9, 877. [Google Scholar] [CrossRef] [Green Version]
  30. Ganal, M.W.; Röder, M.S. Microsatellite and SNP Markers in Wheat Breeding. In Genomics-Assisted Crop Improvement: Vol 2: Genomics Applications in Crops; Varshney, R.K., Tuberosa, R., Eds.; Springer: Dordrecht, The Netherlands, 2007; pp. 1–24. [Google Scholar]
  31. Slate, J.; Gratten, J.; Beraldi, D.; Stapley, J.; Hale, M.; Pemberton, J.M. Gene mapping in the wild with SNPs: Guidelines and future directions. Genetica 2009, 136, 97–107. [Google Scholar] [CrossRef]
  32. Mishra, A.; Singh, P.K.; Bhandawat, A.; Sharma, V.; Sharma, V.; Singh, P.; Roy, J.; Sharma, H. Chapter 8—Analysis of SSR and SNP markers. In Bioinformatics; Singh, D.B., Pathak, R.K., Eds.; Academic Press: Cambridge, MA, USA, 2022; pp. 131–144. [Google Scholar]
  33. Heffner, E.L.; Sorrells, M.E.; Jannink, J.L. Genomic selection for crop improvement. Crop Sci. 2009, 49, 1–12. [Google Scholar] [CrossRef]
  34. Cabrera-Bosquet, L.; Crossa, J.; von Zitzewitz, J.; Serret, M.D.; Luis Araus, J. High-throughput phenotyping and genomic selection: The frontiers of crop breeding converge F. J. Integr. Plant Biol. 2012, 54, 312–320. [Google Scholar] [CrossRef] [Green Version]
  35. Olatoye, M.O.; Clark, L.V.; Wang, J.; Yang, X.; Yamada, T.; Sacks, E.J.; Lipka, A.E. Evaluation of genomic selection and marker-assisted selection in Miscanthus and energycane. Mol. Breed. 2019, 39, 171. [Google Scholar] [CrossRef] [Green Version]
  36. Deomano, E.; Jackson, P.; Wei, X.; Aitken, K.; Kota, R.; Pérez-Rodríguez, P. Genomic prediction of sugar content and cane yield in sugar cane clones in different stages of selection in a breeding program, with and without pedigree information. Mol. Breed. 2020, 40, 38. [Google Scholar] [CrossRef]
  37. Hayes, B.J.; Wei, X.; Joyce, P.; Atkin, F.; Deomano, E.; Yue, J.; Nguyen, L.; Ross, E.M.; Cavallaro, T.; Aitken, K.S. Accuracy of genomic prediction of complex traits in sugarcane. Theor. Appl. Genet. 2021, 134, 1455–1462. [Google Scholar] [CrossRef]
  38. Islam, M.S.; McCord, P.H.; Olatoye, M.O.; Qin, L.; Sood, S.; Lipka, A.E.; Todd, J.R. Experimental evaluation of genomic selection prediction for rust resistance in sugarcane. Plant Genome 2021, 14, e20148. [Google Scholar] [CrossRef]
  39. Andru, S.; Pan, Y.-B.; Thongthawee, S.; Burner, D.M.; Kimbeng, C.A. Genetic analysis of the sugarcane (Saccharum spp.) cultivar ‘LCP 85-384′. I. Linkage mapping using AFLP, SSR, and TRAP markers. Theor. Appl. Genet. 2011, 123, 77–93. [Google Scholar] [CrossRef]
  40. Matsuoka, S.; Ferro, J.; Arruda, P. The Brazilian experience of sugarcane ethanol industry. Vitr. Cell. Dev. Biol.-Plant 2009, 45, 372–381. [Google Scholar] [CrossRef]
  41. Jackson, P.A. Breeding for improved sugar content in sugarcane. Field Crops Res. 2005, 92, 277–290. [Google Scholar] [CrossRef]
  42. Ballesta, P.; Bush, D.; Silva, F.F.; Mora, F.J.P. Genomic predictions using low-density SNP markers, pedigree and GWAS information: A case study with the non-model species Eucalyptus cladocalyx. Plants 2020, 9, 99. [Google Scholar] [CrossRef] [Green Version]
  43. Thirugnanasambandam, P.P.; Singode, A.; Thalambedu, L.; Shanmugavel, S. Opportunity and Challenges for High-throughput Genotyping in Sugarcane. Genotyping By Seq. Crop Improv. 2022, 306–327. [Google Scholar] [CrossRef]
  44. Abramson, M.; Shoseyov, O.; Shani, Z. Plant cell wall reconstruction toward improved lignocellulosic production and processability. Plant Sci. 2010, 178, 61–72. [Google Scholar] [CrossRef]
  45. Liu, P.; Chandra, A.; Que, Y.; Chen, P.-H.; Grisham, M.P.; White, W.H.; Dalley, C.D.; Tew, T.L.; Pan, Y.-B. Identification of quantitative trait loci controlling sucrose content based on an enriched genetic linkage map of sugarcane (Saccharum spp. hybrids) cultivar ‘LCP 85-384′. Euphytica 2016, 207, 527–549. [Google Scholar] [CrossRef]
  46. Islam, M.S.; Pan, Y.B.; Lomax, L.; Grisham, M.P. Identification of quantitative trait loci (QTL) controlling fibre content of sugarcane (Saccharum hybrids spp.). Plant Breed. 2021, 140, 360–366. [Google Scholar] [CrossRef]
  47. Flint-Garcia, S.A.; Thornsberry, J.M.; Buckler IV, E.S. Structure of linkage disequilibrium in plants. Annu. Rev. Plant Biol 2003, 54, 357–374. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  48. Chatterjee, N.; Wheeler, B.; Sampson, J.; Hartge, P.; Chanock, S.J.; Park, J.-H. Projecting the performance of risk prediction based on polygenic analyses of genome-wide association studies. Nat. Genet. 2013, 45, 400–405. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  49. Thomas, D. Methods for investigating gene-environment interactions in candidate pathway and genome-wide association studies. Annu. Rev. Public Health 2010, 31, 21. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  50. Korte, A.; Farlow, A. The advantages and limitations of trait analysis with GWAS: A review. Plant Methods 2013, 9, 29. [Google Scholar] [CrossRef] [Green Version]
  51. Yadav, S.; Jackson, P.; Wei, X.; Ross, E.M.; Aitken, K.; Deomano, E.; Atkin, F.; Hayes, B.J.; Voss-Fels, K.P. Accelerating genetic gain in sugarcane breeding using genomic selection. Agronomy 2020, 10, 585. [Google Scholar] [CrossRef] [Green Version]
  52. Sandhu, K.S.; Shiv, A.; Kaur, G.; Meena, M.R.; Raja, A.K.; Vengavasi, K.; Mall, A.K.; Kumar, S.; Singh, P.K.; Singh, J. Integrated Approach in Genomic Selection to Accelerate Genetic Gain in Sugarcane. Plants 2022, 11, 2139. [Google Scholar] [CrossRef]
  53. Alwala, S.; Kimbeng, C.; Gravois, K.; Bischoff, K. TRAP, a new tool for sugarcane breeding: Comparison with AFLP and coefficient of parentage. J. Am. Soc. Sugar Cane Technol. 2006, 26, 62–86. [Google Scholar]
  54. Creste, S.; Sansoli, D.; Tardiani, A.; Silva, D.; Gonçalves, F.; Fávero, T.; Medeiros, C.; Festucci, C.; Carlini-Garcia, L.; Landell, M. Comparison of AFLP, TRAP and SSRs in the estimation of genetic relationships in sugarcane. Sugar Tech 2010, 12, 150–154. [Google Scholar] [CrossRef]
  55. Legendre, B.; Henderson, M. History and development of sugar yield calculations. Proc.-Am. Soc. Sugar Cane Technol. 1973, 2, 10–18. [Google Scholar]
  56. Aitken, K.; Jackson, P.; McIntyre, C. Quantitative trait loci identified for sugar related traits in a sugarcane (Saccharum spp.) cultivar× Saccharum officinarum population. Theor. Appl. Genet. 2006, 112, 1306–1317. [Google Scholar] [CrossRef]
  57. Pan, Y.-B.; Burner, D.; Legendre, B. An assessment of the phylogenetic relationship among sugarcane and related taxa based on the nucleotide sequence of 5S rRNA intergenic spacers. Genetica 2000, 108, 285–295. [Google Scholar] [CrossRef]
  58. Pan, Y.; Scheffler, B.; Richard Jr, E. High-throughput molecular genotyping of commercial sugarcane clones with microsatellite (SSR) markers. Sugar Tech 2007, 9, 176–181. [Google Scholar]
  59. Pritchard, J.K.; Stephens, M.; Donnelly, P. Inference of population structure using multilocus genotype data. Genetics 2000, 155, 945–959. [Google Scholar] [CrossRef] [PubMed]
  60. Evanno, G.; Regnaut, S.; Goudet, J. Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Mol. Ecol. 2005, 14, 2611–2620. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  61. Earl, D.A.; VonHoldt, B.M. STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv. Genet. Resour. 2012, 4, 359–361. [Google Scholar] [CrossRef]
  62. Shi, A.; Qin, J.; Mou, B.; Correll, J.; Weng, Y.; Brenner, D.; Feng, C.; Motes, D.; Yang, W.; Dong, L. Genetic diversity and population structure analysis of spinach by single-nucleotide polymorphisms identified through genotyping-by-sequencing. PLoS ONE 2017, 12, e0188745. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  63. Kumar, S.; Stecher, G.; Tamura, K. MEGA7: Molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 2016, 33, 1870–1874. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  64. Bradbury, P.J.; Zhang, Z.; Kroon, D.E.; Casstevens, T.M.; Ramdoss, Y.; Buckler, E. TASSEL: Software for association mapping of complex traits in diverse samples. Bioinformatics 2007, 23, 2633–2635. [Google Scholar] [CrossRef]
  65. Jiang, J.; Nguyen, T. Linear and Generalized Linear Mixed Models and Their Applications; Springer: New York, NY, USA, 2007; Volume 1. [Google Scholar]
  66. Liu, X.; Huang, M.; Fan, B.; Buckler, E.S.; Zhang, Z. Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PLoS Genet. 2016, 12, e1005767. [Google Scholar] [CrossRef]
  67. Lipka, A.E.; Tian, F.; Wang, Q.; Peiffer, J.; Li, M.; Bradbury, P.J.; Gore, M.A.; Buckler, E.S.; Zhang, Z. GAPIT: Genome association and prediction integrated tool. Bioinformatics 2012, 28, 2397–2399. [Google Scholar] [CrossRef] [Green Version]
  68. Endelman, J.B. Ridge regression and other kernels for genomic selection with R package rrBLUP. Plant Genome 2011, 4, 250–255. [Google Scholar] [CrossRef] [Green Version]
  69. Heslot, N.; Yang, H.P.; Sorrells, M.E.; Jannink, J.L. Genomic selection in plant breeding: A comparison of models. Crop Sci. 2012, 52, 146–160. [Google Scholar] [CrossRef]
  70. Shikha, M.; Kanika, A.; Rao, A.R.; Mallikarjuna, M.G.; Gupta, H.S.; Nepolean, T. Genomic selection for drought tolerance using genome-wide SNPs in maize. Front. Plant Sci. 2017, 8, 550. [Google Scholar] [CrossRef] [PubMed] [Green Version]
Plants 12 01041 g001 550
Figure 1. Combined violin boxplots of fiber and sucrose contents (%) over two years among 237 clones in the LCP 85-384 population: fiber contents in 2006 (blue) (A) and 2007 (red) (B), sucrose contents in 2006 (green) (C) and 2007 (yellow) (D).
Figure 1. Combined violin boxplots of fiber and sucrose contents (%) over two years among 237 clones in the LCP 85-384 population: fiber contents in 2006 (blue) (A) and 2007 (red) (B), sucrose contents in 2006 (green) (C) and 2007 (yellow) (D).
Plants 12 01041 g001
Plants 12 01041 g002 550
Figure 2. Structural analysis of 237 clones in the LCP 85-384 population based on 135 polymorphic markers (SSR, AFLP and TRAP). (A): Delta K values for different numbers of populations assumed (K = 10) in the STRUCTURE analysis. (B): Classification of 237 clones in three groups (K = 3) using STRUCTURE. The distribution of accessions to different sub-populations is color coded. The X-axis represent the 237 clones from the LCP 85-384 population and the value on the y-axis shows the likelihood of every individual belonging to one of the three colored subpopulations.
Figure 2. Structural analysis of 237 clones in the LCP 85-384 population based on 135 polymorphic markers (SSR, AFLP and TRAP). (A): Delta K values for different numbers of populations assumed (K = 10) in the STRUCTURE analysis. (B): Classification of 237 clones in three groups (K = 3) using STRUCTURE. The distribution of accessions to different sub-populations is color coded. The X-axis represent the 237 clones from the LCP 85-384 population and the value on the y-axis shows the likelihood of every individual belonging to one of the three colored subpopulations.
Plants 12 01041 g002
Plants 12 01041 g003 550
Figure 3. The phylogenetic and principal component analysis of 237 clones in the LCP 85-384 population based on 135 polymorphic markers (SSR, AFLP and TRAP). (A): Phylogenetic analysis of the 237 clones with the corresponding labels as Q groups according to the Structure results: Q1 = red pie, Q2 = green square, Q3 = blue triangle. (B): Scatter diagram of PCA for 237 clones in three color-coded Q groups (A) and mixed Q groups in gray color.
Figure 3. The phylogenetic and principal component analysis of 237 clones in the LCP 85-384 population based on 135 polymorphic markers (SSR, AFLP and TRAP). (A): Phylogenetic analysis of the 237 clones with the corresponding labels as Q groups according to the Structure results: Q1 = red pie, Q2 = green square, Q3 = blue triangle. (B): Scatter diagram of PCA for 237 clones in three color-coded Q groups (A) and mixed Q groups in gray color.
Plants 12 01041 g003
Plants 12 01041 g004 550
Figure 4. Graphs showing QQ-plots for sucrose and fiber contents collected in 2006 and 2007 by four GWAS models: SMR, GLM, MLM, and FarmCPU. Red Square: sucrose in 2006, Blue Pie: sucrose in 2007, Green Triangle: Mean sucrose 2006 and 2007; Yellow Diamond: fiber in 2006, Purple Rectangle: fiber in 2007, Cyan Inverted triangle: Mean fiber in 2006 and 2007.
Figure 4. Graphs showing QQ-plots for sucrose and fiber contents collected in 2006 and 2007 by four GWAS models: SMR, GLM, MLM, and FarmCPU. Red Square: sucrose in 2006, Blue Pie: sucrose in 2007, Green Triangle: Mean sucrose 2006 and 2007; Yellow Diamond: fiber in 2006, Purple Rectangle: fiber in 2007, Cyan Inverted triangle: Mean fiber in 2006 and 2007.
Plants 12 01041 g004
Plants 12 01041 g005 550
Figure 5. Genomic prediction (GP) accuracy (r-value) for fiber (FIB) and sucrose (SUC) using five GP models: Ridge regression best linear unbiased predictor (rrBLUP) = purple, Bayes ridge regression (BRR) = blue; Bayes A (BA) = red, Bayes B (BB) = dark yellow, and Bayesian least absolute shrinkage and selection operator (BL) = green, based on All-allele set and Trait-associated allele set.
Figure 5. Genomic prediction (GP) accuracy (r-value) for fiber (FIB) and sucrose (SUC) using five GP models: Ridge regression best linear unbiased predictor (rrBLUP) = purple, Bayes ridge regression (BRR) = blue; Bayes A (BA) = red, Bayes B (BB) = dark yellow, and Bayesian least absolute shrinkage and selection operator (BL) = green, based on All-allele set and Trait-associated allele set.
Plants 12 01041 g005
Table
Table 1. Phenotypic data indicating mean, standard deviation (Std Dev), standard error (Std Err), min, max, range, variance, coefficient of variation (CV), median, broad-sense heritability (h2) and correlation coefficient (r) of sucrose and fiber contents in 2006 and 2007 among 237 clones in the LCP 85-384 population.
Table 1. Phenotypic data indicating mean, standard deviation (Std Dev), standard error (Std Err), min, max, range, variance, coefficient of variation (CV), median, broad-sense heritability (h2) and correlation coefficient (r) of sucrose and fiber contents in 2006 and 2007 among 237 clones in the LCP 85-384 population.
TraitsYearMean
(%)		Std DevStd ErrMin
(%)		Max
(%)		Range (%)VarianceCV (%)Median (%)Broad-Sense Heritability (h2)Correlations (r)
SucroseFiber
Sucrose200612.981.020.079.8415.545.701.047.8713.070.651.00−0.40
200711.921.230.086.2514.197.941.5110.3012.040.671.00−0.23
Fiber200619.421.890.1314.9624.179.213.589.7519.270.74−0.401.00
200719.381.980.1314.8426.0211.183.9210.2219.440.77−0.231.00
Table
Table 2. The molecular markers associated with sucrose and fiber contents in LCP 85-384 population based on four GWAS models: single marker regression (SMR), generalized linear model (GLM), mixed linear model (MLM) and fixed and random model circulating probability unification (FarmCPU) with data of 2006, 2007 and mean (2006 and 2007), where only LOD [−log(p-value)] > 2.0 were listed.
Table 2. The molecular markers associated with sucrose and fiber contents in LCP 85-384 population based on four GWAS models: single marker regression (SMR), generalized linear model (GLM), mixed linear model (MLM) and fixed and random model circulating probability unification (FarmCPU) with data of 2006, 2007 and mean (2006 and 2007), where only LOD [−log(p-value)] > 2.0 were listed.
MarkerLOD [−log(p-Value)]Trait
SMR GLM MLM FarmCPUSMR GLM MLM FarmCPUSMR GLM MLM FarmCPU
20062007Mean (2006 and 2007)
E32M49_110 3.313.303.053.053.663.593.383.44Sucrose content
E32M50_2902.332.592.422.44 2.892.842.462.71
E32M61_246 2.562.412.262.362.232.202.052.09
E32M62_762.232.462.042.14 3.293.212.503.10
E37M49_1832.292.582.772.20 2.302.20 2.452.822.922.30
E39M50_82 2.032.122.04
SMC703BS_214 2.352.482.272.172.072.092.03
SMC703BS_216 2.352.482.272.172.072.092.03
StSy-R3-128 2.172.212.13 2.232.232.18
E32M49_431 2.452.22 2.362.682.40 Fiber Content
E32M61_1273.672.712.112.702.602.08 2.253.843.092.193.11
E33M61_972.982.942.332.202.952.972.632.553.323.372.592.70
E33M62_1702.662.44 2.322.17 2.003.323.162.192.70
E36M48_614.183.552.893.08 3.282.952.402.66
E36M60_2502.332.11 2.262.13 2.952.842.622.39
E36M61_73 3.263.283.542.812.102.132.27
E37M50_3162.893.072.192.132.792.77 2.413.813.902.723.09
E38M61_1653.773.362.622.78 3.373.192.192.73
E40M59_188 2.25 2.082.372.06
E40M62 153 2.192.40 2.072.17
E41M61_215 2.182.16 3.133.132.742.54
SMC1814LA_152 2.09 2.372.18
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xiong, H.; Chen, Y.; Pan, Y.-B.; Shi, A. A Genome-Wide Association Study and Genomic Prediction for Fiber and Sucrose Contents in a Mapping Population of LCP 85-384 Sugarcane. Plants 2023, 12, 1041. https://doi.org/10.3390/plants12051041

AMA Style

Xiong H, Chen Y, Pan Y-B, Shi A. A Genome-Wide Association Study and Genomic Prediction for Fiber and Sucrose Contents in a Mapping Population of LCP 85-384 Sugarcane. Plants. 2023; 12(5):1041. https://doi.org/10.3390/plants12051041

Chicago/Turabian Style

Xiong, Haizheng, Yilin Chen, Yong-Bao Pan, and Ainong Shi. 2023. "A Genome-Wide Association Study and Genomic Prediction for Fiber and Sucrose Contents in a Mapping Population of LCP 85-384 Sugarcane" Plants 12, no. 5: 1041. https://doi.org/10.3390/plants12051041

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop