Genetic dissection of growth, wood basic density and gene expression in interspecific backcrosses of Eucalyptus grandis and E. urophylla

Haplotype mining panel for genetic dissection and breeding in Eucalyptus

To improve our understanding of genetic mechanisms underlying complex traits in plants, a comprehensive analysis of gene variants is required. Eucalyptus is an important forest plantation genus that is highly outbred. Trait dissection and molecular breeding in eucalypts currently relies on biallelic single-nucleotide polymorphism (SNP) markers. These markers fail to capture the large amount of haplotype diversity in these species, and thus multi-allelic markers are required. We aimed to develop a gene-based haplotype mining panel for Eucalyptus species. We generated 17 999 oligonucleotide probe sets for targeted sequencing of selected regions of 6293 genes implicated in growth and wood properties, pest and disease resistance, and abiotic stress responses. We identified and phased 195 834 SNPs using a read-based phasing approach to reveal SNP-based haplotypes. A total of 8915 target regions (at 4637 gene loci) passed tests for Mendelian inheritance. We evaluated the haplotype panel in four Eucalyptus species (E. grandis, E. urophylla, E. dunnii and E. nitens) to determine its ability to capture diversity across eucalypt species. This revealed an average of 3.13-4.52 haplotypes per target region in each species, and 33.36% of the identified haplotypes were shared by at least two species. This haplotype mining panel will enable the analysis of haplotype diversity within and between species, and provide multi-allelic markers that can be used for genome-wide association studies and gene-based breeding approaches.

Genomic consequences of artificial selection during early domestication of a wood fibre crop

From its origins in Australia, Eucalyptus grandis has spread to every continent, except Antarctica, as a wood crop. It has been cultivated and bred for over 100 yr in places such as South Africa. Unlike most annual crops and fruit trees, domestication of E. grandis is still in its infancy, representing a unique opportunity to interrogate the genomic consequences of artificial selection early in the domestication process. To determine how a century of artificial selection has changed the genome of E. grandis, we generated single nucleotide polymorphism genotypes for 1080 individuals from three advanced South African breeding programmes using the EUChip60K chip, and investigated population structure and genome-wide differentiation patterns relative to wild progenitors. Breeding and wild populations appeared genetically distinct. We found genomic evidence of evolutionary processes known to have occurred in other plant domesticates, including interspecific introgression and intraspecific infusion from wild material. Furthermore, we found genomic regions with increased linkage disequilibrium and genetic differentiation, putatively representing early soft sweeps of selection. This is, to our knowledge, the first study of genomic signatures of domestication in a timber species looking beyond the first few generations of cultivation. Our findings highlight the importance of intra- and interspecific hybridization during early domestication.

Transcriptome analysis of Eucalyptus grandis genotypes reveals constitutive overexpression of genes related to rust (Austropuccinia psidii) resistance

Key Message A resistant E. grandis genotype showed a constitutive overexpression of genes related to resistance to myrtle rust caused by A. psidii. Abstract Myrtle rust caused by Austropuccinia psidii is considered one of the most important fungal diseases affecting Eucalyptus spp. plantations in Brazil. Although the selection and planting of resistant eucalypt genotypes have been the major strategies to manage the disease in Brazil, the molecular mechanisms involved in resistance are still unclear. In this study, we evaluated the gene expression profile of two contrasting Eucalyptus grandis genotypes in resistance level to rust by RNA-Seq. The two genotypes showed a very different background gene expression level even without A. psidii infection. The resistant genotype had a constitutive overexpression of a large number of protein-coding genes compared to the susceptible genotype. These genes were mainly associated with signal transduction, photosynthesis, regulation and response to salicylic acid (SA), and protein kinase leucine-rich receptors (PK-LRR). PK-LRR and SA mediated disease resistance are well known to be effective against obligate biotroph pathogens, such as A. psidii. In addition, at 24 h after infection, the susceptible genotype was able to activate some response, however, several resistance-related proteins had their expression level reduced with A. psidii infection. Here, we present the first analysis of E. grandis genotypes transcriptomes infected by A. psidii and it reveals a constitutive overexpression of several resistance-related genes in the resistant genotype compared to the susceptible one. Our findings have the potential to be used as candidate molecular markers for resistance to myrtle rust.

A systems genetics analysis in Eucalyptus reveals coordination of metabolic pathways associated with xylan modification in wood-forming tissues

Acetyl- and methylglucuronic acid decorations of xylan, the dominant hemicellulose in secondary cell walls (SCWs) of woody dicots, affect its interaction with cellulose and lignin to determine SCW structure and extractability. Genes and pathways involved in these modifications may be targets for genetic engineering; however, little is known about the regulation of xylan modifications in woody plants. To address this, we assessed genetic and gene expression variation associated with xylan modification in developing xylem of Eucalyptus grandis × Eucalyptus urophylla interspecific hybrids. Expression quantitative trait locus (eQTL) mapping identified potential regulatory polymorphisms affecting gene expression modules associated with xylan modification. We identified 14 putative xylan modification genes that are members of five expression modules sharing seven trans-eQTL hotspots. The xylan modification genes are prevalent in two expression modules. The first comprises nucleotide sugar interconversion pathways supplying the essential precursors for cellulose and xylan biosynthesis. The second contains genes responsible for phenylalanine biosynthesis and S-adenosylmethionine biosynthesis required for glucuronic acid and monolignol methylation. Co-expression and co-regulation analyses also identified four metabolic sources of acetyl coenxyme A that appear to be transcriptionally coordinated with xylan modification. Our systems genetics analysis may provide new avenues for metabolic engineering to alter wood SCW biology for enhanced biomass processability.

RNA-seq analysis of lignocellulose-related genes in hybrid Eucalyptus with contrasting wood basic density

BACKGROUND

Wood basic density (WBD), the biomass of plant cell walls per unit volume, is an important trait for elite tree selection in kraft pulp production. Here, we investigated the correlation between WBD and wood volumes or wood properties using 98 open-pollinated, 2.4 to 2.8 year-old hybrid Eucalyptus (Eucalyptus urophylla x E. grandis). Transcript levels of lignocellulose biosynthesis-related genes were studied.

RESULTS

The progeny plants had average WBD of 516 kg/m³ with normal distribution and did not show any correlations between WBD and wood volume or components of α-cellulose, hemicellulose and Klason lignin content. Transcriptomic analysis of two groups of five plants each with high (570-609 kg/m³) or low (378-409 kg/m³) WBD was carried out by RNA-Seq analysis with total RNAs extracted from developing xylem tissues at a breast height. Lignocellulose biosynthesis-related genes, such as cellulose synthase, invertase, cinnamate-4-hydroxylase and cinnamoyl-CoA reductase showed higher transcript levels in the high WBD group. Among plant cell wall modifying genes, increased transcript levels of several expansin and xyloglucan endo-transglycosylase/hydrolase genes were also found in high WBD plants. Interestingly, strong transcript levels of several cytoskeleton genes encoding tubulin, actin and myosin were observed in high WBD plants. Furthermore, we also found elevated transcript levels of genes encoding NAC, MYB, basic helix-loop-helix, homeodomain, WRKY and LIM transcription factors in the high WBD plants. All these results indicate that the high WBD in plants has been associated with the increased transcription of many genes related to lignocellulose formation.

CONCLUSIONS

Most lignocellulose biosynthesis related genes exhibited a tendency to transcribe at relatively higher level in high WBD plants. These results suggest that lignocellulose biosynthesis-related genes may be associated with WBD.

Thirteen years under arid conditions: exploring marker-trait associations in Eucalyptus cladocalyx for complex traits related to flowering, stem form and growth

We present an association analysis for seven key traits related to flowering, stem form and growth in Eucalyptus cladocalyx, a tree species suitable for low rainfall sites, using a long-term progeny trial with 49 open-pollinated maternal families in the southern Atacama Desert, Chile. The progeny trial was carried out in an arid environment with a mean annual rainfall of 152 mm. Simple sequence repeats (SSR) from a full consensus map of Eucalyptus were used for genotyping 245 individual trees. Twenty-three significant marker-trait associations were identified, explaining between 5.9 and 23.7% of the phenotypic variance. The marker EMBRA101 located on LG10 at 56.5 cM was concomitantly associated with diameter at breast height and tree height. Nine SSR were significantly associated with stem forking and stem straightness, explaining between 5.9 and 14.8% of the phenotypic variation. To our knowledge, this is the first study reporting a SSR-based association mapping analysis for stem form traits in Eucalyptus. These results provide novel and valuable information for understanding the genetic base of key traits in E. cladocalyx for breeding purposes under arid conditions.

De Novo Transcriptome Study Identifies Candidate Genes Involved in Resistance to Austropuccinia psidii (Myrtle Rust) in Syzygium luehmannii (Riberry)

Austropuccinia psidii, causal agent of myrtle rust, was discovered in Australia in 2010 and has since become established on a wide range of species within the family Myrtaceae. Syzygium luehmannii, endemic to Australia, is an increasingly valuable berry crop. Plants were screened for responses to A. psidii inoculation, and specific resistance, in the form of localized necrosis, was determined in 29% of individuals. To understand the molecular basis underlying this response, mRNA was sequenced from leaf samples taken preinoculation, and at 24 and 48 h postinoculation, from four resistant and four susceptible plants. Analyses, based on de novo transcriptome assemblies for all plants, identified significant expression changes in resistant plants (438 transcripts) 48 h after pathogen exposure compared with susceptible plants (three transcripts). Most significantly up-regulated in resistant plants were gene homologs for transcription factors, receptor-like kinases, and enzymes involved in secondary metabolite pathways. A putative G-type lectin receptor-like kinase was exclusively expressed in resistant individuals and two transcripts incorporating toll/interleukin-1, nucleotide binding site, and leucine-rich repeat domains were up-regulated in resistant plants. The results of this study provide the first early gene expression profiles for a plant of the family Myrtaceae in response to the myrtle rust pathogen.

Gene co-expression network connectivity is an important determinant of selective constraint

While several studies have investigated general properties of the genetic architecture of natural variation in gene expression, few of these have considered natural, outbreeding populations. In parallel, systems biology has established that a general feature of biological networks is that they are scale-free, rendering them buffered against random mutations. To date, few studies have attempted to examine the relationship between the selective processes acting to maintain natural variation of gene expression and the associated co-expression network structure. Here we utilised RNA-Sequencing to assay gene expression in winter buds undergoing bud flush in a natural population of Populus tremula, an outbreeding forest tree species. We performed expression Quantitative Trait Locus (eQTL) mapping and identified 164,290 significant eQTLs associating 6,241 unique genes (eGenes) with 147,419 unique SNPs (eSNPs). We found approximately four times as many local as distant eQTLs, with local eQTLs having significantly higher effect sizes. eQTLs were primarily located in regulatory regions of genes (UTRs or flanking regions), regardless of whether they were local or distant. We used the gene expression data to infer a co-expression network and investigated the relationship between network topology, the genetic architecture of gene expression and signatures of selection. Within the co-expression network, eGenes were underrepresented in network module cores (hubs) and overrepresented in the periphery of the network, with a negative correlation between eQTL effect size and network connectivity. We additionally found that module core genes have experienced stronger selective constraint on coding and non-coding sequence, with connectivity associated with signatures of selection. Our integrated genetics and genomics results suggest that purifying selection is the primary mechanism underlying the genetic architecture of natural variation in gene expression assayed in flushing leaf buds of P. tremula and that connectivity within the co-expression network is linked to the strength of purifying selection.

Network-based integration of systems genetics data reveals pathways associated with lignocellulosic biomass accumulation and processing

As a consequence of their remarkable adaptability, fast growth, and superior wood properties, eucalypt tree plantations have emerged as key renewable feedstocks (over 20 million ha globally) for the production of pulp, paper, bioenergy, and other lignocellulosic products. However, most biomass properties such as growth, wood density, and wood chemistry are complex traits that are hard to improve in long-lived perennials. Systems genetics, a process of harnessing multiple levels of component trait information (e.g., transcript, protein, and metabolite variation) in populations that vary in complex traits, has proven effective for dissecting the genetics and biology of such traits. We have applied a network-based data integration (NBDI) method for a systems-level analysis of genes, processes and pathways underlying biomass and bioenergy-related traits using a segregating Eucalyptus hybrid population. We show that the integrative approach can link biologically meaningful sets of genes to complex traits and at the same time reveal the molecular basis of trait variation. Gene sets identified for related woody biomass traits were found to share regulatory loci, cluster in network neighborhoods, and exhibit enrichment for molecular functions such as xylan metabolism and cell wall development. These findings offer a framework for identifying the molecular underpinnings of complex biomass and bioprocessing-related traits. A more thorough understanding of the molecular basis of plant biomass traits should provide additional opportunities for the establishment of a sustainable bio-based economy.

The Eucalyptus grandis NBS-LRR Gene Family: Physical Clustering and Expression Hotspots

Eucalyptus grandis is a commercially important hardwood species and is known to be susceptible to a number of pests and pathogens. Determining mechanisms of defense is therefore a research priority. The published genome for E. grandis has aided the identification of one important class of resistance (R) genes that incorporate nucleotide binding sites and leucine-rich repeat domains (NBS-LRR). Using an iterative search process we identified NBS-LRR gene models within the E. grandis genome. We characterized the gene models and identified their genomic arrangement. The gene expression patterns were examined in E. grandis clones, challenged with a fungal pathogen (Chrysoporthe austroafricana) and insect pest (Leptocybe invasa). One thousand two hundred and fifteen putative NBS-LRR coding sequences were located which aligned into two large classes, Toll or interleukin-1 receptor (TIR) and coiled-coil (CC) based on NB-ARC domains. NBS-LRR gene-rich regions were identified with 76% organized in clusters of three or more genes. A further 272 putative incomplete resistance genes were also identified. We determined that E. grandis has a higher ratio of TIR to CC classed genes compared to other woody plant species as well as a smaller percentage of single NBS-LRR genes. Transcriptome profiles indicated expression hotspots, within physical clusters, including expression of many incomplete genes. The clustering of putative NBS-LRR genes correlates with differential expression responses in resistant and susceptible plants indicating functional relevance for the physical arrangement of this gene family. This analysis of the repertoire and expression of E. grandis putative NBS-LRR genes provides an important resource for the identification of novel and functional R-genes; a key objective for strategies to enhance resilience.

Comparative Genomics Analyses Reveal Extensive Chromosome Colinearity and Novel Quantitative Trait Loci in Eucalyptus

Dense genetic maps, along with quantitative trait loci (QTLs) detected on such maps, are powerful tools for genomics and molecular breeding studies. In the important woody genus Eucalyptus, the recent release of E. grandis genome sequence allows for sequence-based genomic comparison and searching for positional candidate genes within QTL regions. Here, dense genetic maps were constructed for E. urophylla and E. tereticornis using genomic simple sequence repeats (SSR), expressed sequence tag (EST) derived SSR, EST-derived cleaved amplified polymorphic sequence (EST-CAPS), and diversity arrays technology (DArT) markers. The E. urophylla and E. tereticornis maps comprised 700 and 585 markers across 11 linkage groups, totaling at 1,208.2 and 1,241.4 cM in length, respectively. Extensive synteny and colinearity were observed as compared to three earlier DArT-based eucalypt maps (two maps with E. grandis × E. urophylla and one map of E. globulus) and with the E. grandis genome sequence. Fifty-three QTLs for growth (10-56 months of age) and wood density (56 months) were identified in 22 discrete regions on both maps, in which only one colocalizaiton was found between growth and wood density. Novel QTLs were revealed as compared with those previously detected on DArT-based maps for similar ages in Eucalyptus. Eleven to 585 positional candidate genes were obained for a 56-month-old QTL through aligning QTL confidence interval with the E. grandis genome. These results will assist in comparative genomics studies, targeted gene characterization, and marker-assisted selection in Eucalyptus and the related taxa.

Transcriptionally active LTR retrotransposons in Eucalyptus genus are differentially expressed and insertionally polymorphic

BACKGROUND

In Eucalyptus genus, studies on genome composition and transposable elements (TEs) are particularly scarce. Nearly half of the recently released Eucalyptus grandis genome is composed by retrotransposons and this data provides an important opportunity to understand TE dynamics in Eucalyptus genome and transcriptome.

RESULTS

We characterized nine families of transcriptionally active LTR retrotransposons from Copia and Gypsy superfamilies in Eucalyptus grandis genome and we depicted genomic distribution and copy number in two Eucalyptus species. We also evaluated genomic polymorphism and transcriptional profile in three organs of five Eucalyptus species. We observed contrasting genomic and transcriptional behavior in the same family among different species. RLC_egMax_1 was the most prevalent family and RLC_egAngela_1 was the family with the lowest copy number. Most families of both superfamilies have their insertions occurring <3 million years, except one Copia family, RLC_egBianca_1. Protein theoretical models suggest different properties between Copia and Gypsy domains. IRAP and REMAP markers suggested genomic polymorphisms among Eucalyptus species. Using EST analysis and qRT-PCRs, we observed transcriptional activity in several tissues and in all evaluated species. In some families, osmotic stress increases transcript values.

CONCLUSION

Our strategy was successful in isolating transcriptionally active retrotransposons in Eucalyptus, and each family has a particular genomic and transcriptional pattern. Overall, our results show that retrotransposon activity have differentially affected genome and transcriptome among Eucalyptus species.

Eucalyptus urograndis stem proteome is responsive to short-term cold stress

Eucalyptus urograndis is a hybrid eucalyptus of major economic importance to the Brazilian pulp and paper industry. Although widely used in forest nurseries around the country, little is known about the biochemical changes imposed by environmental stress in this species. In this study, we evaluated the changes in the stem proteome after short-term stimulation by exposure to low temperature. Using two-dimensional gel electrophoresis coupled to high-resolution mass spectrometry-based protein identification, 12 proteins were found to be differentially regulated and successfully identified after stringent database searches against a protein database from a closely related species (Eucalyptus grandis). The identification of these proteins indicated that the E. urograndis stem proteome responded quickly to low temperature, mostly by down-regulating specific proteins involved in energy metabolism, protein synthesis and signaling. The results of this study represent the first step in understanding the molecular and biochemical responses of E. urograndis to thermal stress.

Microsatellite resources of Eucalyptus: current status and future perspectives

Eucalyptus is the premier paper pulp, short rotation plantation species grown all over the world. Genetic improvement programs integrating molecular marker tools are in progress in many parts of the globe to increase the productivity. Whole genome sequence and expressed sequence tags (ESTs) of the eucalypts paved way for introduction of molecular genetics and breeding in this genus. Different molecular characterization approaches have been used simultaneously in eucalypts, however, microsatellites or simple sequence repeats (SSRs) with their prolific characteristics could occupy a special niche in Eucalyptus genetic improvement. Further, highly informative SSRs were used for the clonal identity, genetic fidelity and in certification of breeder's rights. Eucalyptus genetic linkage maps generated with microsatellite loci were used successfully to identify quantitative trait loci (QTLs) for various economically important traits. Progressively more numbers of microsatellites are being linked to genes associated with adaptive and functional variations, therefore making their utility broader in genetic applications. Availability of common SSR markers across the species provides an opportunity to validate the expression of QTLs across variable genetic backgrounds and accurately compare the position of QTLs in other species. Recent evidences suggest that the presence of SSRs in micro RNAs of plant species play a role in the quantitative trait expression. Similar studies in eucalypts may provide new insights into the genetic architecture of transcript-level variations and post transcriptional gene regulation. This review on eucalypts microsatellites, highlights the availability and characteristics of genomic and eSSRs and their potential in genetic analysis of natural and breeding populations and also discusses the future prospects in population genetics and marker assisted selection.

Cell wall-related proteins of unknown function: missing links in plant cell wall development

Lignocellulosic biomass is an important feedstock for the pulp and paper industry as well as emerging biofuel and biomaterial industries. However, the recalcitrance of the secondary cell wall to chemical or enzymatic degradation remains a major hurdle for efficient extraction of economically important biopolymers such as cellulose. It has been estimated that approximately 10-15% of about 27,000 protein-coding genes in the Arabidopsis genome are dedicated to cell wall development; however, only about 130 Arabidopsis genes thus far have experimental evidence validating cell wall function. While many genes have been implicated through co-expression analysis with known genes, a large number are broadly classified as proteins of unknown function (PUFs). Recently the functionality of some of these unknown proteins in cell wall development has been revealed using reverse genetic approaches. Given the large number of cell wall-related PUFs, how do we approach and subsequently prioritize the investigation of such unknown genes that may be essential to or influence plant cell wall development and structure? Here, we address the aforementioned question in two parts; we first identify the different kinds of PUFs based on known and predicted features such as protein domains. Knowledge of inherent features of PUFs may allow for functional inference and a concomitant link to biological context. Secondly, we discuss omics-based technologies and approaches that are helping identify and prioritize cell wall-related PUFs by functional association. In this way, hypothesis-driven experiments can be designed for functional elucidation of many proteins that remain missing links in our understanding of plant cell wall biosynthesis.

Biochemical and molecular changes associated with heteroxylan biosynthesis in Neolamarckia cadamba (Rubiaceae) during xylogenesis

Wood, derived from plant secondary growth, is a commercially important material. Both cellulose and lignin assembly have been well studied during wood formation (xylogenesis), but heteroxylan biosynthesis is less well defined. Elucidation of the heteroxylan biosynthetic pathway is crucial to understand the mechanism of wood formation. Here, we use Neolamarckia cadamba, a fast-growing tropical tree, as a sample to analyze heteroxylan formation at the biochemical and molecular levels during wood formation. Analysis of the non-cellulosic polysaccharides isolated from N. cadamba stems shows that heteroxylans dominate non-cellulosic polysaccharides and increase with xylogenesis. Microsomes isolated from stems of 1-year-old N. cadamba exhibited UDP-Xyl synthase and xylosyltransferase activities with the highest activity present in the middle and basal stem regions. To further understand the genetic basis of heteroxylan synthesis, RNA sequencing (RNA-seq) was used to generate transcriptomes of N. cadamba during xylogenesis. The RNA-seq results showed that genes related to heteroxylan synthesis had higher expression levels in the middle and basal part of the stem compared to the apical part. Our results describe the heteroxylan distribution and heteroxylan synthesis trait in N. cadamba and give a new example for understanding the mechanism of heteroxylan synthesis in tropical tree species in future.

Impacts of population structure and analytical models in genome-wide association studies of complex traits in forest trees: a case study in Eucalyptus globulus

The promise of association genetics to identify genes or genomic regions controlling complex traits has generated a flurry of interest. Such phenotype-genotype associations could be useful to accelerate tree breeding cycles, increase precision and selection intensity for late expressing, low heritability traits. However, the prospects of association genetics in highly heterozygous undomesticated forest trees can be severely impacted by the presence of cryptic population and pedigree structure. To investigate how to better account for this, we compared the GLM and five combinations of the Unified Mixed Model (UMM) on data of a low-density genome-wide association study for growth and wood property traits carried out in a Eucalyptus globulus population (n = 303) with 7,680 Diversity Array Technology (DArT) markers. Model comparisons were based on the degree of deviation from the uniform distribution and estimates of the mean square differences between the observed and expected p-values of all significant marker-trait associations detected. Our analysis revealed the presence of population and family structure. There was not a single best model for all traits. Striking differences in detection power and accuracy were observed among the different models especially when population structure was not accounted for. The UMM method was the best and produced superior results when compared to GLM for all traits. Following stringent correction for false discoveries, 18 marker-trait associations were detected, 16 for tree diameter growth and two for lignin monomer composition (S∶G ratio), a key wood property trait. The two DArT markers associated with S∶G ratio on chromosome 10, physically map within 1 Mbp of the ferulate 5-hydroxylase (F5H) gene, providing a putative independent validation of this marker-trait association. This study details the merit of collectively integrate population structure and relatedness in association analyses in undomesticated, highly heterozygous forest trees, and provides additional insights into the nature of complex quantitative traits in Eucalyptus.

Plasticity of primary and secondary growth dynamics in Eucalyptus hybrids: a quantitative genetics and QTL mapping perspective

BACKGROUND

The genetic basis of growth traits has been widely studied in forest trees. Quantitative trait locus (QTL) studies have highlighted the presence of both stable and unstable genomic regions accounting for biomass production with respect to tree age and genetic background, but results remain scarce regarding the interplay between QTLs and the environment. In this study, our main objective was to dissect the genetic architecture of the growth trajectory with emphasis on genotype x environment interaction by measuring primary and secondary growth covering intervals connected with environmental variations.

RESULTS

Three different trials with the same family of Eucalyptus urophylla x E. grandis hybrids (with different genotypes) were planted in the Republic of Congo, corresponding to two QTL mapping experiments and one clonal test. Height and radial growths were monitored at regular intervals from the seedling stage to five years old. The correlation between growth increments and an aridity index revealed that growth before two years old (r = 0.5; 0.69) was more responsive to changes in water availability than late growth (r = 0.39; 0.42) for both height and circumference. We found a regular increase in heritability with time for cumulative growth for both height [0.06 - 0.33] and circumference [0.06 - 0.38]. Heritabilities for incremental growth were more heterogeneous over time even if ranges of variation were similar (height [0-0.31]; circumference [0.19 to 0.48]). Within the trials, QTL analysis revealed collocations between primary and secondary growth QTLs as well as between early growth increments and final growth QTLs. Between trials, few common QTLs were detected highlighting a strong environmental effect on the genetic architecture of growth, validated by significant QTL x E interactions.

CONCLUSION

These results suggest that early growth responses to water availability determine the genetic architecture of total growth at the mature stage and highlight the importance of considering growth as a composite trait (such as yields for annual plants) for a better understanding of its genetic bases.