Enhancing genomic prediction with Stacking Ensemble Learning in Arabica Coffee.

Coffee Breeding programs have traditionally relied on observing plant characteristics over years, a slow and costly process. Genomic selection (GS) offers a DNA-based alternative for faster selection of superior cultivars. Stacking Ensemble Learning (SEL) combines multiple models for potentially even more accurate selection. This study explores SEL potential in coffee breeding, aiming to improve prediction accuracy for important traits [yield (YL), total number of the fruits (NF), leaf miner infestation (LM), and cercosporiosis incidence (Cer)] in Coffea Arabica. We analyzed data from 195 individuals genotyped for 21,211 single-nucleotide polymorphism (SNP) markers. To comprehensively assess model performance, we employed a cross-validation (CV) scheme. Genomic Best Linear Unbiased Prediction (GBLUP), multivariate adaptive regression splines (MARS), Quantile Random Forest (QRF), and Random Forest (RF) served as base learners. For the meta-learner within the SEL framework, various options were explored, including Ridge Regression, RF, GBLUP, and Single Average. The SEL method was able to predict the predictive ability (PA) of important traits in Coffea Arabica. SEL presented higher PA compared with those obtained for all base learner methods. The gains in PA in relation to GBLUP were 87.44% (the ratio between the PA obtained from best Stacking model and the GBLUP), 37.83%, 199.82%, and 14.59% for YL, NF, LM and Cer, respectively. Overall, SEL presents a promising approach for GS. By combining predictions from multiple models, SEL can potentially enhance the PA of GS for complex traits.

Identification of SNP Markers and Candidate Genes Associated with Major Agronomic Traits in Coffea arabica.

Genome-wide association studies (GWASs) allow for inferences about the relationships between genomic variants and phenotypic traits in natural or breeding populations. However, few have used this methodology in Coffea arabica. We aimed to identify chromosomal regions with significant associations between SNP markers and agronomic traits in C. arabica. We used a coffee panel consisting of 195 plants derived from 13 families in F2 generations and backcrosses of crosses between leaf rust-susceptible and -resistant genotypes. The plants were phenotyped for 18 agronomic markers and genotyped for 21,211 SNP markers. A GWAS enabled the identification of 110 SNPs with significant associations (p < 0.05) for several agronomic traits in C. arabica: plant height, plagiotropic branch length, number of vegetative nodes, canopy diameter, fruit size, cercosporiosis incidence, and rust incidence. The effects of each SNP marker associated with the traits were analyzed, such that they can be used for molecular marker-assisted selection. For the first time, a GWAS was used for these important agronomic traits in C. arabica, enabling applications in accelerated coffee breeding through marker-assisted selection and ensuring greater efficiency and time reduction. Furthermore, our findings provide preliminary knowledge to further confirm the genomic loci and potential candidate genes contributing to various structural and disease-related traits of C. arabica.

Full-sib progenies show greater genetic diversity than half-sib progenies in sour passion fruit: an approach by ssr markers.

BACKGROUND: Genetic variability is the most important parameter in plant breeding based on selection. There is a need for morpho-agronomic and molecular characterization of Passiflora species, to exploit their genetic resources more efficiently. No study has yet been carried out to compare half-sib and full-sib families in relation to the magnitude of the genetic variability obtained in them, and then to elucidate the advantages or disadvantages of each one. METHODS AND RESULTS: In the present study, SSR markers were used to evaluate the genetic structure and diversity of half-sib and full-sib progenies of sour passion fruit. Two full-sib progenies (PSA and PSB), and a half-sib progeny (PHS), together with their parents, were genotyped with a set of eight pairs of SSR markers. Discriminant Analysis of Principal Components (DAPC) and Structure software were used to study the genetic structure of the progenies. The results indicate that the half-sib progeny has lower genetic variability, although it has higher allele richness. By the AMOVA most of the genetic variability was found within the progenies. Three groups were clearly observed in the DAPC analysis, while two hypothetical groups (k = 2) were observed in the Bayesian approach. The PSB progeny showed a high genetic mixture between the PSA and PHS progenies. CONCLUSION: Lower genetic variability is found in half-sib progenies. The results obtained here allow us to suppose that the selection within full-sib progenies will possibly provide better estimates of genetic variance in sour passion fruit breeding programs, since they provide greater genetic diversity.

Marker-Assisted Recurrent Selection for Pyramiding Leaf Rust and Coffee Berry Disease Resistance Alleles in Coffea arabica L.

In this study, marker-assisted recurrent selection was evaluated for pyramiding resistance gene alleles against coffee leaf rust (CLR) and coffee berry diseases (CBD) in Coffea arabica. A total of 144 genotypes corresponding to 12 hybrid populations from crosses between eight parent plants with desired morphological and agronomic traits were evaluated. Molecular data were used for cross-certification, diversity study and resistance allele marker-assisted selection (MAS) against the causal agent of coffee leaf rust (Hemileia vastatrix) and coffee berry disease (Colletotrichum kahawae). In addition, nine morphological and agronomic traits were evaluated to determine the components of variance, select superior hybrids, and estimate genetic gain. From the genotypes evaluated, 134 were confirmed as hybrids. The genetic diversity between and within populations was 75.5% and 24.5%, respectively, and the cluster analysis revealed three primary groups. Pyramiding of CLR and CBD resistance genes was conducted in 11 genotypes using MAS. A selection intensity of 30% resulted in a gain of over 50% compared to the original population. Selected hybrids with increased gain also showed greater genetic divergence in addition to the pyramided resistance alleles. The strategies used were, therefore, efficient to select superior coffee hybrids for recurrent selection programs and could be used as a source of resistance in various crosses.

Designing the best breeding strategy for Coffea canephora: Genetic evaluation of pure and hybrid individuals aiming to select for productivity and disease resistance traits.

Breeding programs of the species Coffea canephora rely heavily on the significant genetic variability between and within its two varietal groups (conilon and robusta). The use of hybrid families and individuals has been less common. The objectives of this study were to evaluate parents and families from the populations of conilon, robusta, and its hybrids and to define the best breeding and selection strategies for productivity and disease resistance traits. As such, 71 conilon clones, 56 robusta clones, and 20 hybrid families were evaluated over several years for the following traits: vegetative vigor, incidence of rust and cercosporiosis, fruit ripening time, fruit size, plant height, canopy diameter, and yield per plant. Components of variance and genetic parameters were estimated via residual maximum likelihood (REML) and genotypic values were predicted via best linear unbiased prediction (BLUP). Genetic variability among parents (clones) and hybrid families was detected for most of the evaluated traits. The Mulamba-Rank index suggests potential gains up to 17% for the genotypic aggregate of traits in the hybrid population. An intrapopulation recurrent selection within the hybrid population would be the best breeding strategy because the genetic variability, narrow and broad senses heritabilities and selective accuracies for important traits were maximized in the crossed population. Besides, such strategy is simple, low cost and quicker than the concurrent reciprocal recurrent selection in the two parental populations, and this maximizes the genetic gain for unit of time.

Genomic prediction of leaf rust resistance to Arabica coffee using machine learning algorithms

Genomic selection (GS) emphasizes the simultaneous prediction of the genetic effects of thousands of scattered markers over the genome. Several statistical methodologies have been used in GS for the prediction of genetic merit. In general, such methodologies require certain assumptions about the data, such as the normality of the distribution of phenotypic values. To circumvent the non-normality of phenotypic values, the literature suggests the use of Bayesian Generalized Linear Regression (GBLASSO). Another alternative is the models based on machine learning, represented by methodologies such as Artificial Neural Networks (ANN), Decision Trees (DT) and related possible refinements such as Bagging, Random Forest and Boosting. This study aimed to use DT and its refinements for predicting resistance to orange rust in Arabica coffee. Additionally, DT and its refinements were used to identify the importance of markers related to the characteristic of interest. The results were compared with those from GBLASSO and ANN. Data on coffee rust resistance of 245 Arabica coffee plants genotyped for 137 markers were used. The DT refinements presented equal or inferior values of Apparent Error Rate compared to those obtained by DT, GBLASSO, and ANN. Moreover, DT refinements were able to identify important markers for the characteristic of interest. Out of 14 of the most important markers analyzed in each methodology, 9.3 markers on average were in regions of quantitative trait loci (QTLs) related to resistance to disease listed in the literature.(AU)

Receptor-Like Kinase (RLK) as a candidate gene conferring resistance to Hemileia vastatrix in coffee

The biotrophic fungus Hemileia vastatrix causes coffee leaf rust (CLR), one of the most devastating diseases in Coffea arabica. Coffee, like other plants, has developed effective mechanisms to recognize and respond to infections caused by pathogens. Plant resistance gene analogs (RGAs) have been identified in certain plants as candidates for resistance (R) genes or membrane receptors that activate the R genes. The RGAs identified in different plants possess conserved domains that play specific roles in the fight against pathogens. Despite the importance of RGAs, in coffee plants these genes and other molecular mechanisms of disease resistance are still unknown. This study aimed to sequence and characterize candidate genes from coffee plants with the potential for involvement in resistance to H. vastatrix . Sequencing was performed based on a library of bacterial artificial chromosomes (BAC) of the coffee clone Híbrido de Timor (HdT) CIFC 832/2 and screened using a functional marker. Two RGAs, HdT_LRR_RLK1 and HdT_LRR_RLK2, containing the motif of leucine-rich repeat-like kinase (LRR-RLK) were identified. Based on the presence or absence of the HdT_LRR_RLK2 RGA in a number of differential coffee clones containing different combinations of the rust resistance gene, these RGAs did not correspond to any resistance gene already characterized (SH1-9). These genes were also analyzed using qPCR and demonstrated a major expression peak at 24 h after inoculation in both the compatible and incompatible interactions between coffee and H. vastatrix . These results are valuable information for breeding programs aimed at developing CLR-resistant cultivars, in addition to enabling a better understanding of the interactions between coffee and H. vastatrix .(AU)

Genomic prediction of leaf rust resistance to Arabica coffee using machine learning algorithms

Genomic selection (GS) emphasizes the simultaneous prediction of the genetic effects of thousands of scattered markers over the genome. Several statistical methodologies have been used in GS for the prediction of genetic merit. In general, such methodologies require certain assumptions about the data, such as the normality of the distribution of phenotypic values. To circumvent the non-normality of phenotypic values, the literature suggests the use of Bayesian Generalized Linear Regression (GBLASSO). Another alternative is the models based on machine learning, represented by methodologies such as Artificial Neural Networks (ANN), Decision Trees (DT) and related possible refinements such as Bagging, Random Forest and Boosting. This study aimed to use DT and its refinements for predicting resistance to orange rust in Arabica coffee. Additionally, DT and its refinements were used to identify the importance of markers related to the characteristic of interest. The results were compared with those from GBLASSO and ANN. Data on coffee rust resistance of 245 Arabica coffee plants genotyped for 137 markers were used. The DT refinements presented equal or inferior values of Apparent Error Rate compared to those obtained by DT, GBLASSO, and ANN. Moreover, DT refinements were able to identify important markers for the characteristic of interest. Out of 14 of the most important markers analyzed in each methodology, 9.3 markers on average were in regions of quantitative trait loci (QTLs) related to resistance to disease listed in the literature.

Receptor-Like Kinase (RLK) as a candidate gene conferring resistance to Hemileia vastatrix in coffee

The biotrophic fungus Hemileia vastatrix causes coffee leaf rust (CLR), one of the most devastating diseases in Coffea arabica. Coffee, like other plants, has developed effective mechanisms to recognize and respond to infections caused by pathogens. Plant resistance gene analogs (RGAs) have been identified in certain plants as candidates for resistance (R) genes or membrane receptors that activate the R genes. The RGAs identified in different plants possess conserved domains that play specific roles in the fight against pathogens. Despite the importance of RGAs, in coffee plants these genes and other molecular mechanisms of disease resistance are still unknown. This study aimed to sequence and characterize candidate genes from coffee plants with the potential for involvement in resistance to H. vastatrix . Sequencing was performed based on a library of bacterial artificial chromosomes (BAC) of the coffee clone Híbrido de Timor (HdT) CIFC 832/2 and screened using a functional marker. Two RGAs, HdT_LRR_RLK1 and HdT_LRR_RLK2, containing the motif of leucine-rich repeat-like kinase (LRR-RLK) were identified. Based on the presence or absence of the HdT_LRR_RLK2 RGA in a number of differential coffee clones containing different combinations of the rust resistance gene, these RGAs did not correspond to any resistance gene already characterized (SH1-9). These genes were also analyzed using qPCR and demonstrated a major expression peak at 24 h after inoculation in both the compatible and incompatible interactions between coffee and H. vastatrix . These results are valuable information for breeding programs aimed at developing CLR-resistant cultivars, in addition to enabling a better understanding of the interactions between coffee and H. vastatrix .

In silico guided structural and functional analysis of genes with potential involvement in resistance to coffee leaf rust: A functional marker based approach.

Physiology-based differentiation of SH genes and Hemileia vastatrix races is the principal method employed for the characterization of coffee leaf rust resistance. Based on the gene-for-gene theory, nine major rust resistance genes (SH1-9) have been proposed. However, these genes have not been characterized at the molecular level. Consequently, the lack of molecular data regarding rust resistance genes or candidates is a major bottleneck in coffee breeding. To address this issue, we screened a BAC library with resistance gene analogs (RGAs), identified RGAs, characterized and explored for any SH related candidate genes. Herein, we report the identification and characterization of a gene (gene 11), which shares conserved sequences with other SH genes and displays a characteristic polymorphic allele conferring different resistance phenotypes. Furthermore, comparative analysis of the two RGAs belonging to CC-NBS-LRR revealed more intense diversifying selection in tomato and grape genomes than in coffee. For the first time, the present study has unveiled novel insights into the molecular nature of the SH genes, thereby opening new avenues for coffee rust resistance molecular breeding. The characterized candidate RGA is of particular importance for further biological function analysis in coffee.

Correction to: High throughput transcriptome analysis of coffee reveals prehaustorial resistance in response to Hemileia vastatrix infection.

All the transcriptome sequencing data mentioned in the original article is publicly available at the National Center of Biotechnology Information (NCBI).

Genome sequencing and transcript analysis of Hemileia vastatrix reveal expression dynamics of candidate effectors dependent on host compatibility.

Coffee leaf rust caused by the fungus Hemileia vastatrix is one of the most important leaf diseases of coffee plantations worldwide. Current knowledge of the H. vastatrix genome is limited and only a small fraction of the total fungal secretome has been identified. In order to obtain a more comprehensive understanding of its secretome, we aimed to sequence and assemble the entire H. vastatrix genome using two next-generation sequencing platforms and a hybrid assembly strategy. This resulted in a 547 Mb genome of H. vastatrix race XXXIII (Hv33), with 13,364 predicted genes that encode 13,034 putative proteins with transcriptomic support. Based on this proteome, 615 proteins contain putative secretion peptides, and lack transmembrane domains. From this putative secretome, 111 proteins were identified as candidate effectors (EHv33) unique to H. vastatrix, and a subset consisting of 17 EHv33 genes was selected for a temporal gene expression analysis during infection. Five genes were significantly induced early during an incompatible interaction, indicating their potential role as pre-haustorial effectors possibly recognized by the resistant coffee genotype. Another nine genes were significantly induced after haustorium formation in the compatible interaction. Overall, we suggest that this fungus is able to selectively mount its survival strategy with effectors that depend on the host genotype involved in the infection process.

Early Selection Enabled by the Implementation of Genomic Selection in Coffea arabica Breeding.

Genomic Selection (GS) has allowed the maximization of genetic gains per unit time in several annual and perennial plant species. However, no GS studies have addressed Coffea arabica, the most economically important species of the genus Coffea. Therefore, this study aimed (i) to evaluate the applicability and accuracy of GS in the prediction of the genomic estimated breeding value (GEBV); (ii) to estimate the genetic parameters; and (iii) to evaluate the time reduction of the selection cycle by GS in Arabica coffee breeding. A total of 195 Arabica coffee individuals, belonging to 13 families in generation of F2, susceptible backcross and resistant backcross, were phenotyped for 18 agronomic traits, and genotyped with 21,211 SNP molecular markers. Phenotypic data, measured in 2014, 2015, and 2016, were analyzed by mixed models. GS analyses were performed by the G-BLUP method, using the RKHS (Reproducing Kernel Hilbert Spaces) procedure, with a Bayesian algorithm. Heritabilities and selective accuracies were estimated, revealing moderate to high magnitude for most of the traits evaluated. Results of GS analyses showed the possibility of reducing the cycle time by 50%, maximizing selection gains per unit time. The effect of marker density on GS analyses was evaluated. Genomic selection proved to be promising for C. arabica breeding. The agronomic traits presented high complexity for they are controlled by several QTL and showed low genomic heritabilities, evidencing the need to incorporate genomic selection methodologies to the breeding programs of this species.

High throughput transcriptome analysis of coffee reveals prehaustorial resistance in response to Hemileia vastatrix infection.

KEY MESSAGE: We provide a transcriptional profile of coffee rust interaction and identified putative up regulated resistant genes Coffee rust disease, caused by the fungus Hemileia vastatrix, is one of the major diseases in coffee throughout the world. The use of resistant cultivars is considered to be the most effective control strategy for this disease. To identify candidate genes related to different mechanism defense in coffee, we present a time-course comparative gene expression profile of Caturra (susceptible) and Híbrido de Timor (HdT, resistant) in response to H. vastatrix race XXXIII infection. The main objectives were to obtain a global overview of transcriptome in both interaction, compatible and incompatible, and, specially, analyze up-regulated HdT specific genes with inducible resistant and defense signaling pathways. Using both Coffea canephora as a reference genome and de novo assembly, we obtained 43,159 transcripts. At early infection events (12 and 24 h after infection), HdT responded to the attack of H. vastatrix with a larger number of up-regulated genes than Caturra, which was related to prehaustorial resistance. The genes found in HdT at early hours were involved in receptor-like kinases, response ion fluxes, production of reactive oxygen species, protein phosphorylation, ethylene biosynthesis and callose deposition. We selected 13 up-regulated HdT-exclusive genes to validate by real-time qPCR, which most of them confirmed their higher expression in HdT than in Caturra at early stage of infection. These genes have the potential to assist the development of new coffee rust control strategies. Collectively, our results provide understanding of expression profiles in coffee-H. vastatrix interaction over a time course in susceptible and resistant coffee plants.

Molecular characterization of arabica and Conilon coffee plants genotypes by SSR and ISSR markers

The molecular characterization of ten genotypes of the Coffea arabica plants and of seven genotypes of C. canephora having interesting features for coffee breeding programs was carried to select the parents for breeding. A total of 40 SSR and 29 ISSR primers were used. The primers generated a total of 331 (307 polymorphic and 24 monomorphic) bands. Analysis of genetic diversity presented dissimilarity intervals ranging from 0.22 to 0.44 between the Conilon genotypes, from 0.02 to 0.28 between the Arabica genotypes, and from 0.49 to 0.60 between the genotypes of the two species in the joint analysis. Four groups were formed: I = genotypes of C. arabica, II = four progenies of C. canephora, Conilon group, and one non defined C. canephora (Conilon or Robusta), III = one progeny of un-defined C. canephora (Conilon or Robusta) and IV = one progeny of C. canephora of Robusta group. The grouping formed was consistent with the origins of each group. High stabilities of the bifurcations were found by bootstrap analysis. The use of molecular markers of the SSR and ISSR types in the diversity study was efficient in distinguishing genotypes between and within C. arabica and C. canephora. .

In silico identification of coffee genome expressed sequences potentially associated with resistance to diseases.

Sequences potentially associated with coffee resistance to diseases were identified by in silico analyses using the database of the Brazilian Coffee Genome Project (BCGP). Keywords corresponding to plant resistance mechanisms to pathogens identified in the literature were used as baits for data mining. Expressed sequence tags (ESTs) related to each of these keywords were identified with tools available in the BCGP bioinformatics platform. A total of 11,300 ESTs were mined. These ESTs were clustered and formed 979 EST-contigs with similarities to chitinases, kinases, cytochrome P450 and nucleotide binding site-leucine rich repeat (NBS-LRR) proteins, as well as with proteins related to disease resistance, pathogenesis, hypersensitivity response (HR) and plant defense responses to diseases. The 140 EST-contigs identified through the keyword NBS-LRR were classified according to function. This classification allowed association of the predicted products of EST-contigs with biological processes, including host defense and apoptosis, and with molecular functions such as nucleotide binding and signal transducer activity. Fisher's exact test was used to examine the significance of differences in contig expression between libraries representing the responses to biotic stress challenges and other libraries from the BCGP. This analysis revealed seven contigs highly similar to catalase, chitinase, protein with a BURP domain and unknown proteins. The involvement of these coffee proteins in plant responses to disease is discussed.

In silico identification of coffee genome expressed sequences potentially associated with resistance to diseases

Sequences potentially associated with coffee resistance to diseases were identified by in silico analyses using the database of the Brazilian Coffee Genome Project (BCGP). Keywords corresponding to plant resistance mechanisms to pathogens identified in the literature were used as baits for data mining. Expressed sequence tags (ESTs) related to each of these keywords were identified with tools available in the BCGP bioinformatics platform. A total of 11,300 ESTs were mined. These ESTs were clustered and formed 979 EST-contigs with similarities to chitinases, kinases, cytochrome P450 and nucleotide binding site-leucine rich repeat (NBS-LRR) proteins, as well as with proteins related to disease resistance, pathogenesis, hypersensitivity response (HR) and plant defense responses to diseases. The 140 EST-contigs identified through the keyword NBS-LRR were classified according to function. This classification allowed association of the predicted products of EST-contigs with biological processes, including host defense and apoptosis, and with molecular functions such as nucleotide binding and signal transducer activity. Fisher's exact test was used to examine the significance of differences in contig expression between libraries representing the responses to biotic stress challenges and other libraries from the BCGP. This analysis revealed seven contigs highly similar to catalase, chitinase, protein with a BURP domain and unknown proteins. The involvement of these coffee proteins in plant responses to disease is discussed.