Abstract
Lipids are compounds that play an important role in coffee bean development, contributing to beverage quality. Genome-wide association studies (GWAS) were conducted to pinpoint quantitative trait nucleotides (QTNs) linked to lipid metabolism in Coffea arabica. Genotyping by sequencing (GBS) and phenotyping data from 104 wild C. arabica accessions, Mundo Novo cultivar, and C. arabica var. Typica were utilized. GBS data were aligned to C. arabica Et039 reference genome, and both single-locus and multi-locus GWAS methods were employed. Methods were adjusted for kinship matrix, population structure, and principal component analysis. Of the 19 QTNs identified, 5 showed consistency across different population structure adjustments. The multi-locus methods mrMLM and FarmCPU proved more effective in identifying QTNs associated with lipid content. Four QTNs were situated near seven genes potentially involved in lipid metabolism. Higher frequencies of identified QTNs in accessions with elevated lipid content suggest their utility as markers for coffee plant breeding.
Keywords:
Allotetraploid; coffee; GWAS; SNP markers
INTRODUCTION
Coffee is one of the most popular beverages in the world, and its commercial production is based on two species: Coffea arabica and Coffea canephora. Brazil stands out for being the largest producer of this commodity, which contributes directly to the economy and job creation (CONAB 2023CONAB - Companhia Nacional de Abastecimento2023 Acompanhamento da safra brasileira de café. Available at <Available at https://www.conab.gov.br/info-agro/safras/cafe/boletim-da-safra-de-cafe >. Accessed on April 23, 2023.
https://www.conab.gov.br/info-agro/safra...
). Coffea arabica has higher production, mainly due to its superior cup quality, which adds greater economic value. Cup quality in coffee is due to the composition and combination of several compounds, including chlorogenic acids, caffeine, sugars, diterpenes, and lipids (Scholz et al. 2016Scholz MB, Kitzberger CS, Pagiatto NF, Pereira LF, Davrieux F, Pot D, Charmetant P, Leroy T2016 Chemical composition in wild Ethiopian Arabica coffee accessions. Euphytica 209:429). Lipids are key coffee compounds that play an important role in coffee bean development, contributing to the flavor and aroma of coffee beverages (Sant’Ana et al. 2018Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465).
Genome-wide association studies (GWAS), together with next-generation sequencing technology, have emerged as powerful tools for identifying molecular markers associated with agronomic traits of interest. GWAS can overcome the limitations of traditional genetic linkage mapping, including maps with little refinement and limited parental diversity (Bartoli and Roux 2017Bartoli C, Roux F2017 Genome-Wide Association studies in plant pathosystems: Toward an ecological genomics approach. Frontiers in Plant Science 8:763). GWAS can explore the genetic diversity found in wild crop relatives, offering a higher mapping resolution in comparison with biparental quantitative trait loci (QTLs) experiments and is considered a cost-effective way to detect associations between molecular markers and traits of interest (Korte and Farlow 2013Korte A, Farlow A2013 The advantages and limitations of trait analysis with GWAS: A review. Plant Methods 9:29, Su et al. 2016Su J, Pang C, Wei H, Li L, Liang B, Wang C, Song M, Wang H, Zhao S, Jia X, Mao G, Huang L, Geng D, Wang C, Fan S, Yu S2016 Identification of favorable SNP alleles and candidate genes for traits related to early maturity via GWAS in upland cotton. BMC Genomics 17:687).
Sant’Ana et al. (2018Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465) published the first GWAS in C. arabica and identified single nucleotide polymorphisms (SNPs) associated with the biochemical characteristics of the grain, such as lipids, and the diterpenes cafestol and kahweol. However, it was used only in the reference genome of C. canephora (one of the diploid ancestors of the allotetraploid C. arabica), limiting the number of SNPs used in the association.
In this study, our objectives were to use the same genotyping by sequencing (GBS) and lipid phenotype data from previous work but with the GBS data aligned to the complete C. arabica genome in the association study, aiming to identify novel genomic regions linked to lipid metabolism in C. arabica.
MATERIAL AND METHODS
Plant material
A Coffea arabica collection of 104 wild genotypes from Ethiopia, C. arabica Mundo Novo 38, and C. arabica var. Typica was used. GBS and phenotyping for total lipid content in green grains from this population were previously carried out by Sant’Ana et al. (2018Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465). The GBS data were aligned to C. arabica reference genome Et39 (Arabica Coffee Genome Consortium, Salojärvi et al. 2023Salojärvi J, Rambani A, Yu Z, Guyot R, Strickler S, Lepelley M, Wang C, Rajaraman S, Rastas P, Zheng C, Muñoz DS, Meidanis J, Paschoal AR, Bawin Y, Krabbenhoft T, Wang ZQ, Fleck S, Aussel R, Bellanger L, Charpagne A, Fournier C, Kassam M, Lefebvre G, Métairon S, Moine D, Rigoreau M, Stolte J, Hamon P, Couturon E, Tranchant-Dubreuil C, Mukherjee M, Lan T, Engelhardt J, Stadler P, De Lemos SMC, Suzuki SI, Sumirat U, Man WC, Dauchot N, Orozco-Arias S, Garavito A, Kiwuka C, Musoli P, Nalukenge A, Guichoux E, Reinout H, Smit M, Carretero-Paulet L, Filho OG, Braghini MT, Padilha L, Sera GH, Ruttink T, Henry R, Marraccini P, Van de Peer Y, Andrade A, Domingues D, Giuliano G, Mueller L, Pereira LF, Plaisance S, Poncet V, Rombauts S, Sankoff D, Albert VA, Crouzillat D, de Kochko A, Descombes P2023 The genome and population genomics of allopolyploid Coffea arabica reveal the diversification history of modern coffee cultivars. BioRxiv 2023.09.06.556570.) and subjected to SNP calling using the Tassel 5 GBS v2 pipeline, resulting in a panel of 159,000 SNPs (Glaubitz et al. 2014Glaubitz JC, Casstevens TM, Lu F, Harriman J, Elshire RJ, Sun Q, Buckler ES2014 TASSEL-GBS: A high capacity genotyping by sequencing analysis pipeline. PLoS One 9:e90346).
SNP filtering
The 159,000 SNPs aligned to the C. arabica genome were filtered using TASSEL software version 5.2.89, as described by Bradbury et al. (2007Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES2007 TASSEL: Software for association mapping of complex traits in diverse samples. Bioinformatics 23:2633-2635). Filtering was performed with the parameters of minor allele frequency (MAF > 0.05) and call rate > 0.8. For data imputation, Beagle v4.1 software (Browning and Browning 2016Browning BL, Browning SR2016 Genotype imputation with millions of reference samples. The American Journal of Human Genetics 98:116-12) and LD-kNNi based on the K-nearest neighbor method (Money et al. 2015Money D, Gardner K, Migicovsky Z, Schwaninger H, Zhong GY, Myles S2015 LinkImpute: fast and accurate genotype imputation for nonmodel organisms. G3: Genes, Genomes, Genetics 5:2383-2390) were used.
Population structure
The first five principal component analyses (PCA) and kinship matrix (K) were calculated using TASSEL 5.2.53 and used for single-locus GWAS (SL-GWAS) and multi-locus GWAS (ML-GWAS). The Q matrix (Q) was produced using Structure v2.3.4 software (Pritchard et al. 2000Pritchard JK, Stephens M, Donnelly P2000 Inference of population structure using multilocus genotype data. Genetics 155:945-959). For the Q matrix, the allele frequencies of each K cluster (2-10) were estimated, with 1000 runs as the burn-in period, 1000 runs for the Markov chain Monte Carlo (MCMC), and 10 runs for each K value. The ∆K criterion (Evanno et al. 2005Evanno G, Regnaut S, Goudet J2005 Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14:2611-2620) was used in Structure Harvester software (Earl and Vonholdt 2012Earl DA, Vonholdt BM2012 Structure Harvester: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4:359-361) to estimate the uppermost level of the population structure.
Single-locus and multi-locus GWAS
Single-locus GWAS (SL-GWAS) was performed by TASSEL 5.2.53 with two methods: a general linear model (GLM, Price et al. 2006Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D2006 Principal components analysis corrects for stratification in genome-wide association studies. Nature Genetics 38:904-909) and a mixed linear model (MLM, Yu et al. 2006Yu J, Pressoir G, Briggs WH, Vroh Bi I, Yamasaki M, Doebley JF, McMullen MD, Gaut BS, Nielsen DM, Holland JB, Kresovich S, Buckler ES2006 A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nature Genetics 38:203-208). The association threshold for SL-GWAS was p ≤ 0.05/n, where n is the number of markers. Multi-locus GWAS (ML-GWAS) were performed using five methods: mrMLM (Wang et al. 2016Wang SB, Feng JY, Ren WL, Huang B, Zhou L, Wen YJ, Zhang J, Dunwell JM, Xu S, Zhang YM2016 Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Scientific Reports 6:19444), FASTmrMLM (Tamba and Zhang 2018Tamba CL, Zhang YM2018 A fast mrMLM algorithm for multi-locus genome-wide association studies. Available at <Available at https://api.semanticscholar.org/CorpusID:90293082 >. Accessed on April 15, 2023.
https://api.semanticscholar.org/CorpusID...
), FASTmrEMMA (Wen et al. 2018Wen YJ, Zhang H, Ni YL, Huang B, Zhang J, Feng JY, Wang SB, Dunwell JM, Zhang YM, Wu R2018 Methodological implementation of mixed linear models in multi-locus genome-wide association studies. Briefings in Bioinformatics 19:700-712), ISIS EM-BLASSO (Tamba et al. 2017Tamba CL, Ni YL, Zhang YM2017 Iterative sure independence screening EM-Bayesian LASSO algorithm for multi-locus genome-wide association studies. PLoS Computational Biology 13:e1005357), and FarmCPU (Liu et al. 2016Liu X, Huang M, Fan B, Buckler ES, Zhang Z2016 Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PLoS Genetics 12:e1005957) using R software (R Core Team 2023). For mrMLM, FASTmrMLM, FASTmrEMMA, and ISIS EM-BLASSO methods, in the first step, the critical values p ≤ 0.01, 0.01, 0.005, and 0.01 were used, respectively, for the intermediate result. In the second step, all SNPs selected in the first step were filtered by the multi-locus methods, and the markers with the largest effects that exceeded the LOD score threshold were considered potentially associated SNPs. The critical LOD score threshold was set to 3 for SNPs in the final phase. For FarmCPU, the criterion p ≤ 0.0005 was used.
Linkage disequilibrium analysis and identification of candidate genes
Squared correlation coefficients (r 2) were calculated on sliding windows with 50 adjacent SNPs in TASSEL version 5.2.53, which was used to evaluate linkage disequilibrium (LD) decay. The LD distance value in base pairs (bp) was evaluated by a non-linear regression method, at r 2 = 0.2, using R software (R Core Team 2023R Core Team2023 R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at < Available at https://www.r-project.org />. Accessed on July, 20, 2023.
https://www.r-project.org...
). The search for candidate genes was performed using the C. arabica Et039 genome functional annotation according to the LD threshold upstream and downstream of the associated QTN positions.
RESULTS AND DISCUSSION
SNP filtering and population structure
After quality filters and imputation, of the 159,000 SNPs identified in the SNP calling, 11,136 SNPs were used for population structure adjustments and GWAS. Of this total, 5032 were identified in the canephora subgenome, 4870 in the eugenioides subgenome, and 1234 in chromosome zero. Chromosome zero corresponds to scaffolds that do not have a defined position during genome assembly. This represents a significant improvement concerning our first GWAS work, in which the GBS data were aligned only in the C. canephora genome, and GWAS was performed with only 2587 SNPs (Sant’Ana et al. 2018Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465).
PCA-based population structure analysis (Supplementary Figure S1) and a Q matrix from Structure K=3 (the higher delta K) (Supplementary Figure S2) were used for population structure adjustments to the GWAS methods. In Sant’Ana et al. (2018Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465), the higher delta K obtained was K=2 but, interestingly, the second highest value K=3 showed a population structure very similar to the present work. Both the PCA and structure analysis (Supplementary Figure 3) presented here have similar results to those of Ariyoshi et al. (2022Ariyoshi C, Sant’ana GC, Felicio MS, Sera GH, Nogueira LM, Rodrigues LMR, Ferreira RV, Silva BSR, Resende MLV, Deste’fano SAL, Domingues DS, Pereira LFP2022 Genome-wide association study for resistance to Pseudomonas syringae pv. garcae in Coffea arabica. Frontiers in Plant Science 13:989847), who used a similar plant collection and genotyped data mapped in the C. arabica genome.
Single-locus GWAS
Single-locus methods were unable to detect QTNs (Figure 1). A common feature of SL-GWAS is a one-dimensional genome scan, in which each marker is tested in turn. However, this approach does not facilitate good estimates of the effects of markers controlled by multiple loci, which occur in most complex traits (Wang et al. 2016Wang SB, Feng JY, Ren WL, Huang B, Zhou L, Wen YJ, Zhang J, Dunwell JM, Xu S, Zhang YM2016 Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Scientific Reports 6:19444). Another problem with the method is the issue of multiple test corrections for the threshold value of the significance test. Due to the Bonferroni correction, which is normally very conservative, methods such as GLM and MLM are not efficient in detecting small effective loci of a complex trait (Wang et al. 2016Wang SB, Feng JY, Ren WL, Huang B, Zhou L, Wen YJ, Zhang J, Dunwell JM, Xu S, Zhang YM2016 Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Scientific Reports 6:19444).
Manhattan plots (A-B) of the single-locus genome-wide association study for lipid content using the GLM and MLM methods. The x-axis represents the chromosomes, and the y-axis represents the −log10(p-value). The solid line indicates the thresholds for lipid content in C. arabica. Quantile-quantile (Q-Q) plots (C-D) of a single-locus genome-wide association study for lipid content in C. arabica. Q-Q plots show the observed vs. expected negative log10 p-values.
Multi-locus GWAS
In this study, only ML-GWAS detected QTNs related to lipid content. The advantages of ML-GWAS over SL-GWAS have already been described in studies of other plant species, such as cotton, corn, tobacco, and coffee (Li et al. 2018Li C, Fu Y, Sun R, Wang Y, Wang Q2018 Single-locus and multi-locus genome-wide association studies in the genetic dissection of fiber quality traits in upland cotton (Gossypium hirsutum L.). Frontiers in Plant Science 9:1083, Su et al. 2018Su J, Ma Q, Li M, Hao F, Wang C2018 Multi-Locus Genome-Wide Association studies of fiber-quality related traits in Chinese early-maturity upland cotton. Frontiers in Plant Science 9:1169, Xu et al. 2018Xu Y, Yang T, Zhou Y, Yin S, Li P, Liu J, Xu S, Yang Z, Xu C2018 Genome-Wide Association mapping of starch pasting properties in maize using single-locus and multi-locus models. Frontiers in Plant Science 9:1311, Ariyoshi et al. 2022Ariyoshi C, Sant’ana GC, Felicio MS, Sera GH, Nogueira LM, Rodrigues LMR, Ferreira RV, Silva BSR, Resende MLV, Deste’fano SAL, Domingues DS, Pereira LFP2022 Genome-wide association study for resistance to Pseudomonas syringae pv. garcae in Coffea arabica. Frontiers in Plant Science 13:989847, Ikram et al. 2022Ikram M, Xiao J, Li R, Xia Y, Zhao W, Yuan Q, Siddique KHM, Guo P2022 Identification of superior haplotypes and candidate genes for yield-related traits in tobacco (Nicotiana tabacum L.) using association mapping. Industrial Crops and Products 189:115886). This is because multi-locus methods are characterized by a multidimensional genome scanning approach in which the effects of all markers are estimated simultaneously (Cui et al. 2018Cui Y, Zhang F, Zhou Y2018 The application of multi-locus GWAS for the detection of salt-tolerance loci in rice. Frontiers in Plant Science 9:1464).
In the population structure correction by PCA + K (Figure 2), 13 QTNs were identified (Table 1). However, except for the FASTmrEMMA and FarmCPU, the Q-Q plot graphs (Figure 2) depicted observed p-values consistently lower than expected throughout the plot with the PCA + K models. In those methods, the inclusion of PCA as a population structure correction resulted in unfavorable adjustments. In a study developed by Elhaik (2022Elhaik E2022 Principal component analyses (PCA)-based findings in population genetic studies are highly biased and must be reevaluated. Scientific Reports 12:14683), 12 common test cases were analyzed in human population data, in which the adjustment with PCA demonstrated unfavorable results in association studies for the studied trait.
Manhattan plots (A-E) of the multi-locus genome-wide association study for lipid content with the methods mrMLM, FASTmrMLM, FASTmrEMMA, ISIS EM-BLASSO, and FarmCPU with the PCA + K models. The x-axis represents the chromosomes, and the y-axis represents the −log10(p-value). The dashed lines indicate the LOD score threshold and the solid line indicates the −log10(p-value) threshold for lipid content in C. arabica. Quantile-quantile (Q-Q) plots (F-J) of a multi-locus genome-wide association study for lipid content in C. arabica with the PCA + K models. Q-Q plots show the observed vs. expected negative log10 p-values.
Manhattan plots (A-D) of the multi-locus genome-wide association study for lipid content with the methods mrMLM, FASTmrMLM, ISIS EM-BLASSO, and FarmCPU with the Q + K models. The x-axis represents the chromosomes, and the y-axis represents the −log10(p-value). The dashed lines indicate the LOD score threshold, and the solid line indicates the −log10(p-value) threshold, for lipid content in C. arabica. Quantile-quantile (Q-Q) plots (E-H) of a multi-locus genome-wide association study for lipid content in C. arabica with the Q + K models. Q-Q plots show the observed vs. expected negative log10 p-values.
In the adjustment of multi-locus methods, using clustering coefficient data obtained by the Structure software and K matrix (Q + K) (Figure 3), 6 QTNs were identified (Table 2). The methods presented a better adjustment, as shown in the Q-Q plot graphs (Figure 3), except for the FASTmrEMMA method, in which the PCA correction presented a better adjustment of the p-value data but did not associate QTN with the correction by the Q + K models. The inclusion of the Q matrix in the GWAS methods decreased the number of QTNs by approximately three times. Yang et al. (2011Yang X, Gao S, Xu S, Zhang Z, Prasanna BM, Li L, Li J, Yan J2011 Characterization of a global germplasm collection and its potential utilization for analysis of complex quantitative traits in maize. Molecular Breeding 28:511-526) evaluated complex traits in maize using the PCA + K and Q + K models as population structure corrections, in which Q + K showed a better reduction in false positives.
GWAS using PCA + K resulted in 13 QTNs, and in 3 of them (Chr_7_sg_C_409622, Chr_8_sg_C_19869722, and Chr_5_sg_E_35714509), we observed nearby genes related to lipid biosynthesis and/or metabolism. Meanwhile, the GWAS using Q + K resulted in 6 QTNs, 2 of which (Chr_6_sg_C_4567047 and Chr_8_sg_C_19869722) had nearby genes related to lipid genes. The different population structure correction models identified five convergent QTNs (Chr_0_4634_168274, Chr_8_sg_C_19869722, Chr_2_sg_E_60056603, Chr_7_sg_E_20849219, and Chr_11_sg_E_29236360).
Our previous GWAS study related to lipid content identified 5 QTNs (Sant’Ana et al. 2018Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465). In this work, 14 QTNs were identified by the GWAS methods using the PCA + K and Q + K models, a number approximately three times greater than our initial results. Of the 6 QTNs of the canephora subgenome, 4 were on the same pseudo-chromosome as our original work (Sant’Ana et al. 2018).
Identification of candidate genes related to lipid content
In the four QTNs, genes involved in lipid metabolism and/or fatty acid biosynthesis were identified. From the functional annotation of C. arabica Et039, seven genes were identified close to the QTNs associated with the lipid trait (Table 3). To consider a gene linked to a QTN, the approximate distance based on the LD decay result was used. The LD decay r 2 = 0.2 was 158,774 bp (Supplementary Figure S4).
QTN Chr_6_sg_C_4567047 is close to the g15.102 gene, with its functional annotation for 4-phosphopantetheinyl transferase isoform X1. For Chr_5_sg_E_35714509, the nearby gene is g119.153, with functional annotation for the electron transfer flavo mitochondrial subunit. For these two QTNs, no descriptions were found in the literature for a better understanding of their functions.
QTN Chr_7_sg_C_409622 is close to the g10.29 gene with functional annotation for the protein triacylglycerol lipase-like 1. This protein is involved in acyl-lipid metabolism in Arabidopsis thaliana, as well as in other plant species. The acyl lipid has several functions, including providing the central membrane diffusion barrier that separates cells and subcellular organelles. This function alone encompasses more than 10 classes of membrane lipids, such as phospholipids, galactolipids, and sphingolipids (Li-beisson et al. 2013Li-Beisson Y, Shorrosh B, Beisson F, Andersson MX, Arondel V, Bates PD, Baud S, Bird D, DeBono A, Durrett TP, Franke RB, Graham IA, Katayama K, Kelly AA, Larson T, Markham JE, Miquel M, Molina I, Nishida I, Rowland O, Samuels L, Schmid KM, Wada H, Welti R, Xu C, Zallot R, Ohlrogge J2013 Acyl-lipid metabolism. The Arabidopsis Book 11:e0133). In addition to this QTN, in the annotation of C. arabica Et039, genes g10.11 and g10.8 were also identified, with participation in lipid metabolism, with functional annotations for patatin 3 and patatin 1, respectively.
QTN Chr_8_sg_C_19869722 is close to the g66.10 gene, which has a functional annotation for 3-ketoacyl-synthase 10. This protein contributes to the biosynthesis of cuticular wax and suberin (Lolle et al. 1997Lolle SJ, Berlyn GP, Engstrom EM, Krolikowski KA, Reiter WD, Pruitt RE1997 Developmental regulation of cell interactions in the Arabidopsis fiddlehead-1mutant: A role for the epidermal cell wall and cuticle. Developmental Biology 189:311-321). As precursors of wax compounds, very long-chain fatty acids participate in limiting non-stomatal water loss and preventing pathogen attacks. They are also used as energy storage in seeds and as building blocks for membranes. Twenty-one 3-ketoacyl-CoA synthase genes were identified in the Arabidopsis thaliana genome and expressed in seeds, flowers, and leaves (Joubès et al. 2008Joubès J, Raffaele S, Bourdenx B, Garcia C, Laroche-Traineau J, Moreau P, Domergue F, Lessire R2008 The VLCFA elongase gene family in Arabidopsis thaliana: phylogenetic analysis, 3D modelling and expression profiling. Plant Molecular Biology 67:547-566).
CONCLUSION
Using the C. arabica genome as a reference for the GWAS study for lipids represented a significant improvement from our previous work, where only the C. canephora genome was available. This allowed us to recover a higher number of SNPs, increase the number of QTNs, and identify candidate genes involved with lipids and/or fatty acid biosynthesis. The information generated is also important for the development of markers for breeding and for providing candidate genes for transcriptome analysis to depict lipid biosynthesis in C. arabica.
ACKNOWLEDGMENTS
We acknowledge the Consórcio Pesquisa Café for financial support and the Scholarship for H.V.L.M. (Grant 10.18.20.027.00). INCT Café provided fellowships for C.A., R.F.V., and M.S.F. L.F.P.P. acknowledges CNPq for the research fellowship. Supplementary Figures and Tables may be requested from the corresponding author.
REFERENCES
- Ariyoshi C, Sant’ana GC, Felicio MS, Sera GH, Nogueira LM, Rodrigues LMR, Ferreira RV, Silva BSR, Resende MLV, Deste’fano SAL, Domingues DS, Pereira LFP2022 Genome-wide association study for resistance to Pseudomonas syringae pv. garcae in Coffea arabica. Frontiers in Plant Science 13:989847
- Bartoli C, Roux F2017 Genome-Wide Association studies in plant pathosystems: Toward an ecological genomics approach. Frontiers in Plant Science 8:763
- Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES2007 TASSEL: Software for association mapping of complex traits in diverse samples. Bioinformatics 23:2633-2635
- Browning BL, Browning SR2016 Genotype imputation with millions of reference samples. The American Journal of Human Genetics 98:116-12
- CONAB - Companhia Nacional de Abastecimento2023 Acompanhamento da safra brasileira de café. Available at <Available at https://www.conab.gov.br/info-agro/safras/cafe/boletim-da-safra-de-cafe >. Accessed on April 23, 2023.
» https://www.conab.gov.br/info-agro/safras/cafe/boletim-da-safra-de-cafe - Cui Y, Zhang F, Zhou Y2018 The application of multi-locus GWAS for the detection of salt-tolerance loci in rice. Frontiers in Plant Science 9:1464
- Earl DA, Vonholdt BM2012 Structure Harvester: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources 4:359-361
- Elhaik E2022 Principal component analyses (PCA)-based findings in population genetic studies are highly biased and must be reevaluated. Scientific Reports 12:14683
- Evanno G, Regnaut S, Goudet J2005 Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology 14:2611-2620
- Glaubitz JC, Casstevens TM, Lu F, Harriman J, Elshire RJ, Sun Q, Buckler ES2014 TASSEL-GBS: A high capacity genotyping by sequencing analysis pipeline. PLoS One 9:e90346
- Ikram M, Xiao J, Li R, Xia Y, Zhao W, Yuan Q, Siddique KHM, Guo P2022 Identification of superior haplotypes and candidate genes for yield-related traits in tobacco (Nicotiana tabacum L.) using association mapping. Industrial Crops and Products 189:115886
- Joubès J, Raffaele S, Bourdenx B, Garcia C, Laroche-Traineau J, Moreau P, Domergue F, Lessire R2008 The VLCFA elongase gene family in Arabidopsis thaliana: phylogenetic analysis, 3D modelling and expression profiling. Plant Molecular Biology 67:547-566
- Korte A, Farlow A2013 The advantages and limitations of trait analysis with GWAS: A review. Plant Methods 9:29
- Li C, Fu Y, Sun R, Wang Y, Wang Q2018 Single-locus and multi-locus genome-wide association studies in the genetic dissection of fiber quality traits in upland cotton (Gossypium hirsutum L.). Frontiers in Plant Science 9:1083
- Li-Beisson Y, Shorrosh B, Beisson F, Andersson MX, Arondel V, Bates PD, Baud S, Bird D, DeBono A, Durrett TP, Franke RB, Graham IA, Katayama K, Kelly AA, Larson T, Markham JE, Miquel M, Molina I, Nishida I, Rowland O, Samuels L, Schmid KM, Wada H, Welti R, Xu C, Zallot R, Ohlrogge J2013 Acyl-lipid metabolism. The Arabidopsis Book 11:e0133
- Liu X, Huang M, Fan B, Buckler ES, Zhang Z2016 Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PLoS Genetics 12:e1005957
- Lolle SJ, Berlyn GP, Engstrom EM, Krolikowski KA, Reiter WD, Pruitt RE1997 Developmental regulation of cell interactions in the Arabidopsis fiddlehead-1mutant: A role for the epidermal cell wall and cuticle. Developmental Biology 189:311-321
- Money D, Gardner K, Migicovsky Z, Schwaninger H, Zhong GY, Myles S2015 LinkImpute: fast and accurate genotype imputation for nonmodel organisms. G3: Genes, Genomes, Genetics 5:2383-2390
- Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D2006 Principal components analysis corrects for stratification in genome-wide association studies. Nature Genetics 38:904-909
- Pritchard JK, Stephens M, Donnelly P2000 Inference of population structure using multilocus genotype data. Genetics 155:945-959
- R Core Team2023 R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at < Available at https://www.r-project.org />. Accessed on July, 20, 2023.
» https://www.r-project.org - Salojärvi J, Rambani A, Yu Z, Guyot R, Strickler S, Lepelley M, Wang C, Rajaraman S, Rastas P, Zheng C, Muñoz DS, Meidanis J, Paschoal AR, Bawin Y, Krabbenhoft T, Wang ZQ, Fleck S, Aussel R, Bellanger L, Charpagne A, Fournier C, Kassam M, Lefebvre G, Métairon S, Moine D, Rigoreau M, Stolte J, Hamon P, Couturon E, Tranchant-Dubreuil C, Mukherjee M, Lan T, Engelhardt J, Stadler P, De Lemos SMC, Suzuki SI, Sumirat U, Man WC, Dauchot N, Orozco-Arias S, Garavito A, Kiwuka C, Musoli P, Nalukenge A, Guichoux E, Reinout H, Smit M, Carretero-Paulet L, Filho OG, Braghini MT, Padilha L, Sera GH, Ruttink T, Henry R, Marraccini P, Van de Peer Y, Andrade A, Domingues D, Giuliano G, Mueller L, Pereira LF, Plaisance S, Poncet V, Rombauts S, Sankoff D, Albert VA, Crouzillat D, de Kochko A, Descombes P2023 The genome and population genomics of allopolyploid Coffea arabica reveal the diversification history of modern coffee cultivars. BioRxiv 2023.09.06.556570.
- Sant’Ana GC, Pereira LFP, Pot D, Ivamoto ST, Domingues DS, Ferreira RV, Pagiatto NF, Silva BSR, Nogueira LM, Kitzberger CSG, Scholz MBS, Oliveira FF, Sera GH, Padilha L, Labouisse JP, Guyot R, Charmetant P and Leroy T2018 Genome-wide association study reveals candidate genes influencing lipids and diterpenes contents in Coffea arabica L. Scientific Reports 8:465
- Scholz MB, Kitzberger CS, Pagiatto NF, Pereira LF, Davrieux F, Pot D, Charmetant P, Leroy T2016 Chemical composition in wild Ethiopian Arabica coffee accessions. Euphytica 209:429
- Su J, Ma Q, Li M, Hao F, Wang C2018 Multi-Locus Genome-Wide Association studies of fiber-quality related traits in Chinese early-maturity upland cotton. Frontiers in Plant Science 9:1169
- Su J, Pang C, Wei H, Li L, Liang B, Wang C, Song M, Wang H, Zhao S, Jia X, Mao G, Huang L, Geng D, Wang C, Fan S, Yu S2016 Identification of favorable SNP alleles and candidate genes for traits related to early maturity via GWAS in upland cotton. BMC Genomics 17:687
- Tamba CL, Ni YL, Zhang YM2017 Iterative sure independence screening EM-Bayesian LASSO algorithm for multi-locus genome-wide association studies. PLoS Computational Biology 13:e1005357
- Tamba CL, Zhang YM2018 A fast mrMLM algorithm for multi-locus genome-wide association studies. Available at <Available at https://api.semanticscholar.org/CorpusID:90293082 >. Accessed on April 15, 2023.
» https://api.semanticscholar.org/CorpusID:90293082 - Wang SB, Feng JY, Ren WL, Huang B, Zhou L, Wen YJ, Zhang J, Dunwell JM, Xu S, Zhang YM2016 Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Scientific Reports 6:19444
- Wen YJ, Zhang H, Ni YL, Huang B, Zhang J, Feng JY, Wang SB, Dunwell JM, Zhang YM, Wu R2018 Methodological implementation of mixed linear models in multi-locus genome-wide association studies. Briefings in Bioinformatics 19:700-712
- Xu Y, Yang T, Zhou Y, Yin S, Li P, Liu J, Xu S, Yang Z, Xu C2018 Genome-Wide Association mapping of starch pasting properties in maize using single-locus and multi-locus models. Frontiers in Plant Science 9:1311
- Yang X, Gao S, Xu S, Zhang Z, Prasanna BM, Li L, Li J, Yan J2011 Characterization of a global germplasm collection and its potential utilization for analysis of complex quantitative traits in maize. Molecular Breeding 28:511-526
- Yu J, Pressoir G, Briggs WH, Vroh Bi I, Yamasaki M, Doebley JF, McMullen MD, Gaut BS, Nielsen DM, Holland JB, Kresovich S, Buckler ES2006 A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nature Genetics 38:203-208
Publication Dates
-
Publication in this collection
07 June 2024 -
Date of issue
2024
History
-
Received
18 Dec 2023 -
Accepted
07 Mar 2024 -
Published
10 Apr 2024