Genome-wide genotyping elucidates the geographical diversification and dispersal of the polyploid and clonally propagated yam (Dioscorea alata)

Abstract

Background and aims: Inferring the diffusion history of many human-dispersed species is still not straightforward due to unresolved past human migrations. The centre of diversification and routes of migration of the autopolyploid and clonally propagated greater yam, Dioscorea alata, one of the oldest edible tubers, remain unclear. Here, we address yam demographic and dispersal history using a worldwide sample.

Methods: We characterized genome-wide patterns of genetic variation using genotyping by sequencing 643 greater yam accessions spanning four continents. First, we disentangled the polyploid and clonal components of yam diversity using allele frequency distribution and identity by descent approaches. We then addressed yam geographical origin and diffusion history with a model-based coalescent inferential approach.

Key results: Diploid genotypes were more frequent than triploids and tetraploids worldwide. Genetic diversity was generally low and clonality appeared to be a main factor of diversification. The most likely evolutionary scenario supported an early divergence of mainland Southeast Asian and Pacific gene pools with continuous migration between them. The genetic make-up of triploids and tetraploids suggests that they have originated from these two regions before westward yam migration. The Indian Peninsula gene pool gave origin to the African gene pool, which was later introduced to the Caribbean region.

Conclusions: Our results are congruent with the hypothesis of independent domestication origins of the two main Asian and Pacific gene pools. The low genetic diversity and high clonality observed suggest a strong domestication bottleneck followed by thousands of years of widespread vegetative propagation and polyploidization. Both processes reduced the extent of diversity available for breeding, and this is likely to threaten future adaptation.

Keywords: Dioscorea alata; Clonal propagation; demography; geographical distribution; polyploidy; population genomics; yam.

Fig. 1.

Network showing the genetic relationships between diploid, triploid and tetraploid accessions. The combination of colours and shape represents ploidy levels and geographical origin.

Fig. 2.

(A) Clonal relationships within the 40 multi-locus lineages (MLLs) identified among the 302 Dioscorea alata diploid accessions. Each cluster represents one MLL. Node colours correspond to the geographical origin of the clones. –(B) Geographical distribution of MLLs shared between continents. The total number of MLLs within each continent is reported: Africa 13 (shared seven); Asia (shared four); Caribbean 13 (shared 12) and Pacific 12 (shared eight). (See also Table 1.)

Fig. 3.

Visualization of genetic relationships between the 93 diploid Dioscorea alata accessions from Africa, Asia, Caribbean and Pacific after removing the clones. (A) Principal component analysis depicting the 93 accessions. Squares, Africa; circles; Asia; triangles, Caribbean; and diamonds, Pacific. Accessions are coloured according to their assignment to the four genetic clusters after Admixture analysis. The threshold to assign a genotype to a cluster is set at 80 %. (B) Admixture barplot showing the distribution of the K = 4 genetic clusters. Within each continent, accessions are ordered according to cluster assignment proportions. (C) Map showing the geographical distribution of the 93 accessions in each continent according to their genetic clustering.

Fig. 4.

Demographic scenarios of domestication simulated with FASTSIMCOAL2.6. The split between Mainland South East Asia (MSEA) and the Pacific (Pac) was followed by the split of Indian Peninsula (InP) populations with constant migration between InP and Pac. Scenarios: (A) origin of Africa (Afr) from the Indian Peninsula, (B) origin of Africa from the Pacific, (C) origin of Africa from the Indian Peninsula with gene flow between the Pacific and Africa. N_ANC, ancestral population size. TDIV, time of divergence; TDI1 < TIV2 < TDIV3. GF, gene flow. 0.005 is the migration rate between two populations.