Skip to main content
Dryad logo

Georeferenced phylogenetic analysis of a global collection of wild and cultivated Citrullus species


Achigan-Dako, Enoch Gbenato; Degbey, Herve; Hale, Iago; Blattner, Frank (2021), Georeferenced phylogenetic analysis of a global collection of wild and cultivated Citrullus species, Dryad, Dataset,


The geographical origin of watermelon (Citrullus lanatus) remains debated. While a first hypothesis suggests the center of origin to be west Africa, where a sister endemic species C. mucosospermus thrives, a second hypothesis suggests north-eastern Africa where the white-fleshed Sudanese Kordophan melon is cultivated. In this study, we infer biogeographical and haplotype genealogy for C. lanatus, C. mucosospermus, C. amarus, and C. colocynthis using non-coding cpDNA sequences (trnT-trnL and ndhF-rpl32 regions) from a global collection of 135 accessions. In total, we identified 38 haplotypes in C. lanatus, C. mucosospermus, C. amarus, and C. colocynthis; of these, 21 were found in Africa and 17 appear endemic to the continent. The least diverse species was C. mucosospermus (5 haplotypes) and the most diverse was C. colocynthis (16 haplotypes). Some haplotypes of C. mucosospermus were nearly exclusive to West-Africa, and C. lanatus and C. mucosospermus shared haplotypes that were distinct from those of both C. amarus and C. colocynthis. The results support previous findings C. mucosospermus to be the closest relative to C. lanatus (including subsp. cordophanus). West Africa, as a center of endemism of C. mucosospermus, is an area of interest in the search of the origin of C. lanatus. This calls for further historical and phylogeographical investigations and wider collection of samples in West and North-East Africa.


We used 135 accessions of fCitrullus spp. to generate sequence data based on chloroplast regions. Two regions of the chloroplast genome were amplified using the following primer pairs: 1) ndhF (5'-GAAAGGTATKATCAAYGMATATT-3') and rpl32-R (5'-CCAATATCCCTTYYTTTTCCAA-3'); and 2) trnL(UAG) (5'-CTGCTTCCTAAGAGCAGCCT-3') and trnT(GGU) (5'-CCCTTTTAACTCAGTGGTAG-3'). Parsimony networks were constructed to infer phylogeographical relationships among haplotypes using TCS v1.21 (Clément et al., 2000).

Usage Notes

Data can be added to other sources of information to build a meta analysis or combined with other chloroplast regions to infer phylogeny.


Vavilov-Frankel Fellowship, Award: CONT/08/136/RF

Vavilov-Frankel Fellowship, Award: CONT/08/136/RF