Genetic diversity of the cultivated vanilla in Madagascar

Natural vanillin extracted from vanilla pods is the most common flavour in the world, mainly used for foods, drinks and also in perfumery and cosmetics. The producing countries are far from able to meet the growing demand. Vanilla production is limited to some tropical areas such as Madagascar. More than 80,000 Malagasy families rely directly on vanilla production, which constitutes the major part of exportations of the country. However, the usual position of Madagascar as the “head” of providers in terms of volume and quality is currently threatened due to several issues, the most frequent being instability in quantity and quality of the final product. Fungal attacks have been present in the plantations since before 1932 causing a loss of about one fifth of the production in the 1950s. A breeding program (from 1944 to 2000) has offered promise for a sustainable vanilla crop by demonstrating the possibility of creating resistant vanilla varieties. Unfortunately, the program has been abandoned and since then, neither genetic study nor agronomical improvement has been conducted on the Malagasy germplasm. The genetic diversity of the most cultivated vanilla species – Vanilla planifolia – has been previously reported to be very narrow due to its restricted initial diversity from a handful Mexican cuttings used as unique genetic basis in all areas of introduction. The classical markers such as isozymes, RAPD, RFLP, microsatellites; previously used did not successfully explain the variation within V. planifolia. The aim of the present work was first to evaluate the current genetic resources of the unexplored cultivated vanilla in Madagascar, in the attempt of supporting future conservation and breeding programs. Secondly, in order to address the knowledge gap about vanilla, the study seeked to understand the contribution of space and environment parameters in the genetic variation of the crop. For these purposes, we used the single nucleotide polymorphism (SNP) markers, a high throughput and the most abundant marker across genomes, successfully used in several crop plants with low diversity and recently used in vanilla. A large-scale prospect was conducted to cover as much diversity present in Madagascar as possible. Our sampling included 246 accessions collected from the four major vanilla production regions distributed from north to south and northwest of the country (Manakara, SAVA, Mananara Nord and Ambanja). This sampling covered various environmental conditions for vanilla cultivation in Madagascar. Biological materials were collected from 120 plantations, with 30 vanilla fields from each region, distributed between low, medium, and high altitudes. A total of 46 accessions from a local germplasm collection built from the ancient breeding program was also included. This latter was first used to test the effectiveness of the SNP markers in vanilla genetic study and to choose the genotyping tool that fitted better to vanilla. ddRAD-seq protocol was used to generate DNA sequences data and the three computing tools (Stacks, Bcftools and GATK) most widely used for genotyping were compared trough descriptive genetic statistics, population structure and job running time in SNP calling. Then, the genetic diversity and the structure of the Malagasy germplasm was evaluated using the developed pipeline. Lastly, using redundancy analysis, the contribution of space and environment on the SNP variation was computed based on the geographical coordinates and the information about soil, temperature and precipitation of the accession location from QGIS-based data. In general, the present study showed that SNP is an effective marker that can be used for inter- and intra-species levels of vanilla studies. BCFtools was demonstrated to be appropriate to vanilla SNP genotyping due to its ability to genotype highly heterozygous genomes under a reasonable running time. The developed SNP markers were genotyped with enough depth of coverage and were proportionally distributed among the 14 vanilla chromosomes. The set of targeted accessions were genetically segregated according to their respective taxonomic/phenotypic groups with a number of private alleles in each group. The success of the chosen marker was also underlined by the revealed structuration within species. In particular, our results revealed a clear genetic structure of the Malagasy germplasm and highlighted original varieties, mainly arising from the past breeding program. The 17,948 developed high-quality SNPs distinguished the 246 accessions from the field into five genetic groups – V. planifolia, V. pompona, Big Vanilla and two intermediate groups with a similar phenotype intermediate between the two first groups. Two additional genetic groups were identified in the local collection (Vanilla Banane and Tsivaky). Intra-species genetic structuration was also revealed from the current study. V. planifolia accessions were structured into 3 major genetic groups with high variability (9,281 SNPs). However, the genetic distance between groups was low as previously reported. V. pompona accessions were also grouped according to their provenance in the SAVA region, the genetic differentiation according to altitude was evidenced. Redundancy analysis detected a low contribution of spatial and environmental parameters to the variation of SNPs. Therefore, many other unincluded parameters may contribute to the diversity of SNPs. Our results highlighted the clear genetic structuration of the cultivated vanilla in Madagascar with the presence of original varieties. The data provided through this study offer valuable information for future breeding and management programs.