Current efforts to grow the tropical oilseed crop Jatropha curcas L. economically are hampered by the lack of cultivars and the presence of toxic phorbol esters (PE) within the seeds of most provenances. These PE restrict the conversion of seed cake into animal feed, although naturally occurring ‘nontoxic’ provenances exist which produce seed lacking PE. As an important step towards the development of genetically improved varieties of J. curcas, we constructed a linkage map from four F2 mapping populations. The consensus linkage map contains 502 codominant markers, distributed over 11 linkage groups, with a mean marker density of 1.8 cM per unique locus. Analysis of the inheritance of PE biosynthesis indicated that this is a maternally controlled dominant monogenic trait. This maternal control is due to biosynthesis of the PE occurring only within maternal tissues. The trait segregated 3 : 1 within seeds collected from F2 plants, and QTL analysis revealed that a locus on linkage group 8 was responsible for phorbol ester biosynthesis. By taking advantage of the draft genome assemblies of J. curcas and Ricinus communis (castor), a comparative mapping approach was used to develop additional markers to fine map this mutation within 2.3 cM. The linkage map provides a framework for the dissection of agronomic traits in J. curcas, and the development of improved varieties by marker-assisted breeding. The identification of the locus responsible for PE biosynthesis means that it is now possible to rapidly breed new nontoxic varieties.

BackgroundEconomical cultivation of the oilseed crop Jatropha curcas is currently hampered in part due to the non-availability of purpose-bred cultivars. Although genetic maps and genome sequence data exist for this crop, marker-assisted breeding has not yet been implemented due to a lack of available marker–trait association studies. To identify the location of beneficial alleles for use in plant breeding, we performed quantitative trait loci (QTL) analysis for a number of agronomic traits in two biparental mapping populations.ResultsThe mapping populations segregated for a range of traits contributing to oil yield, including plant height, stem diameter, number of branches, total seeds per plant, 100-seed weight, seed oil content and fatty acid composition. QTL were detected for each of these traits and often over multiple years, with some variation in the phenotypic variance explained between different years. In one of the mapping populations where we recorded vegetative traits, we also observed co-localization of QTL for stem diameter and plant height, which were both overdominant, suggesting a possible locus conferring a pleotropic heterosis effect. By using a candidate gene approach and integrating physical mapping data from a recent high-quality release of the Jatropha genome, we were also able to position a large number of genes involved in the biosynthesis of storage lipids onto the genetic map. By comparing the position of these genes with QTL, we were able to detect a number of genes potentially underlying seed traits, including phosphatidate phosphatase genes.ConclusionsThe QTL we have identified will serve as a useful starting point in the creation of new varieties of J. curcas with improved agronomic performance for seed and oil productivity. Our ability to physically map a significant proportion of the Jatropha genome sequence onto our genetic map could also prove useful in identifying the genes underlying particular traits, allowing more controlled and precise introgression of desirable alleles and permitting the pyramiding or stacking of multiple QTL.Electronic supplementary materialThe online version of this article (doi:10.1186/s13068-015-0326-8) contains supplementary material, which is available to authorized users.