Developing pre-breeding tools for the miracle plant [Synsepalum dulcificum (Schumach & Thonn.) Daniell]: implications for genomic selection strategies optimization.

Abstract

The miracle plant [Synsepalum dulcificum (Schumach & Thonn) Daniell] is an evergreen orphan shrub species originating in West Africa that produces red berries known as miracle fruits. As a unique natural source of miraculin, miracle fruits stand as a singular natural sweetener with huge applications and economic potential. Unfortunately, the lack of systematic breeding history in the species currently hinders the full exploitation of the species potential at both the local and international levels. This study was undertaken with the goal of generating information that can be used to initiate a proper breeding and large-scale cultivation program of the species in its center of origin, West Africa, with a focus on three countries (Ghana, Togo, and the Republic of Benin). The objectives of the study were: i) to determine current management practices and end-users’ preferred traits to identify key breeding traits and formulate sound breeding objectives, ii) to evaluate the natural phenotypic diversity in the species to identify potential parental lines/populations for breeding purposes, iii) to assess genomic diversity and population structure in the species to reshape breeding strategies, and iv) to model the determinants of the local communities’ willingness to invest in the species cultivation to articulate promotion strategies in the species. To assess trait preferences by the miracle plant end-users (farmers, consumers, and processing companies) in the study area, semi-structured interviews and focus-group discussions were held with 300 individual respondents from various socio-cultural backgrounds and one processing company each in Benin and Ghana. The results revealed that farmers in Ghana currently manage the species better than their counterparts in Benin, with men being overall the main owners of the species. Our results offer an impetus for West Africa region-wide cultivar development, as both farmers’ and consumers’ preferences for breeding traits were >80% similar among the six sociolinguistic groups and the three agroecological zones considered in the study area. The preference for breeding traits was 60% similar among farmer, consumer and processing company user groups. Out of the 23 potential breeding traits identified in the study area, fruit size, fruit miraculin content, fruit yielding, early maturity, fruit edible ratio and seed part ratio represent priority breeding traits. Interestingly the traits (fruit size and miraculin content) were consistently ranked as the top two desired traits by all three end-user groups of the species. The classification conducted on 203 individual trees (accessions) sampled from seven populations in the Upper Guinea forest block (Ghana) and the Dahomey Gap (Ghana, Togo, and Benin) block revealed the existence of three natural phenotypic groups in the species. While in terms of population performance, the Volta population in the Dahomey Gap outperformed the six other populations for almost all the end-users’ desired traits assessed. Cluster 3, which encompassed accessions from all

seven populations sampled, presented the best performance based on end-users’ preferences and, therefore, represents a potential source of elite accessions. Overall, variability ranged from low to moderate in traits evaluated [coefficient of variation: 9.65% (fruit length) – 53.34% (tree diameter at ground level] and the inferred core collection points to the Dahomey Gap as a center of diversity of the miracle plant. Despite the expected heterozygosity of HS = 0.14 suggesting a low diversity in the miracle plant, the molecular assessment done on 322 individuals revealed a strong population differentiation in the species (FST =0.69). Populations in the Upper Guinea forest block strongly diverged from those in the Dahomey Gap block, and our data supported the existence of a barrier (which we speculated to be the Lake Volta in Ghana) disrupting connectivity between those two West African rainforest block populations of the miracle plant. Corroborating the findings of the phenotypic assessment, the genome-wide diversity analysis also defined three genetic groups. Conveying a higher resolution to the population structuring, the genomic analysis highlighted an ecoregional structuring of the miracle plant population in the study area. The DNA polymorphism and demographic history analyses indicated the Western Dahomey Gap as the likely centre of origin of the miracle plant, thus refining the previously postulated West Africa origin of the species. The Volta population was confirmed as an exceptional breeding population with regards to its high allelic richness and heterozygosity. Our findings also suggest that sociolinguistic groups in Benin valued the miracle plant more than those in Ghana, to the extent that the species diversity was affected, as fitness organs such as roots and bark were highly targeted for medicinal uses. Nonetheless, these sociolinguistic groups in Benin were also more willing to engage in the miracle plant cultivation compared with those in Ghana. However, for this to be effective, market structuring and improved variety (early fruiting) development combined with awareness raising are needed to accelerate local community engagement in the species cultivation in the entire study area. Collectively, these results advance our knowledge of the miracle plant and form the basis for breeding initiatives in the species. The implications of these findings regarding optimization of a genomic selection strategy for accelerated genetic gains achievement in agronomic and functional traits of this species is discussed.