Taro, Colocasia esculenta (L.) Schott, is an ancient crop of great agricultural, cultural, and biological significance (Chaïr et al., 2016; Matthews and Ghanem, 2021; Lebot and Ivančič, 2022; Tan et al., 2025). It is cultivated widely across the globe—from humid tropical to warm temperate regions—and is a common component of diets in the Pacific Islands, Asia, Africa, and parts of the Americas. Even at the northern edge of taro’s climatic range, in northern China, South Korea, and Japan (East Asia), taro is commonly and traditionally grown and consumed as a local food (Siegmund, 2010; Matthews and Ghanem, 2021). Despite its wide cultivation and high nutritional value, taro has often been regarded as an “orphan crop” in scientific research (Matthews and Ghanem, 2021). There is a clear need for more comprehensive study of taro, given its importance for food security and cultural heritage in many communities and countries, and potential for further development as a crop in diverse physical environments.

Taro belongs to the arum family (Araceae) and is one of several species in the genus Colocasia Schott, with most being wild species known only to taxonomists. Recent phylogenomic studies have begun to clarify taro’s evolutionary relationships, natural distribution as a wild breeding species, and global phylogeography as a clonally propagated crop (Chaïr et al., 2016; Ahmed et al., 2020; Matthews et al., 2024). Plastid DNA (ptDNA; also referred to as chloroplast DNA) analyses of cultivated and wild taros have revealed a surprisingly complex genetic structure: C. esculenta is comprised of at least three major plastid lineages (haplogroups) that broadly correspond to different environmental adaptations and geographies. The CI lineage is found in wetland cultivars and commensal wild taros in tropical to subtropical Southeast Asia and Oceania, CII has been found mainly in dryland cultivated taros in temperate East Asia and other subtropical-to-temperate regions, and CIII has been found exclusively in wild taro populations from Southeast Asia through New Guinea to northern Australia.

In Korea, taro (toran, or t’oran, meaning “earth eggs”) has been cultivated for centuries (Siegmund, 2010), yet it remains a relatively minor crop compared to staple cereals (Kim et al., 2019). Despite taro’s long presence in Korean agriculture, its phylogenetic position and genetic diversity in a global context remain uncharacterized. Previous authors have suggested that taro may have been introduced to Japan via China, Southeast Asia, or the Korean peninsula. However, their studies relied on limited molecular characterization (Matsuda and Nawata, 2002; Huh and Choi, 2003) or included only a single Korean sample (Motohashi et al., 2024), or no Korean samples (Matsuda and Nawata, 2002), thereby limiting the resolution of introduction history and past dispersal routes into and across East Asia.

To begin addressing this knowledge gap, we sequenced and analyzed the complete plastid genomes (plastomes) of four cultivated taro samples collected in Korea, during a pilot field survey carried out in late 2025. The new plastome data provide a genomic basis for distinguishing Korean taro lineages and provide a foundation for future comparative studies on the origins, introduction routes, and genetic composition of cultivated taros in East Asia.

Comparative analyses of these plastome sequences allowed us to evaluate genetic divergence among Korean taro samples and to assess their maternal lineage relationships. Phylogenetic analyses (Fig. 6, Online Supplementary Figs. S1, S2) based on CDS, IGS, and complete plastome alignments consistently recovered the two major cultivated taro plastid lineages, CI and CII, with strong support, in agreement with previous studies (Ahmed et al., 2012, 2013, 2020; Matthews et al., 2024).

Among the four Korean taro samples, three (CESKR01, CESKR02, and CESKR03) were consistently placed in the CII lineage across all analyses (CDS, IGS, and complete plastome). These appear identical to two closely-related reference CII haplotypes from New Zealand and northern Pakistan (Fig. 6), indicating that the Korean CII taros share maternal ancestry with other temperate-region cultivars that may have originated in the subtropical mountains of Southeast Asia (eastern Himalaya) (Matthews 2014, 2023). In contrast, CESKR04 was placed in the CI lineage together with a reference CI genome from southern Japan that is also identified as the CI Type 1 haplotype shared by tropical taros across Asia and the Pacific (Ahmed et al., 2020; Matthews et al., 2024). The growth of CESKR04 in Korea’s temperate climate is notable, particularly as this recently introduced cultivar was able to reach maturity and flower. During our survey, we also observed ornamental taro varieties flowering at Gungnamji, an artificial pond in Buyeo-si, Chungcheongnamdo. These observations raise the possibility that controlled breeding with winter protection may be possible in Korea. In the context of global climate change, continued warming may further facilitate the cultivation and spread of tropical (CI) taro lineages in Korea, potentially contributing to shifts in the composition of locally adapted taro cultivars. Collectively, these findings show that taro cultivated in Korea does not derive from a single domesticated lineage but instead comprises multiple, genetically divergent maternal origins. This result underscores the utility of plastome-based phylogenomic approaches for revealing infraspecific diversity and for informing strategies in conservation, breeding, and climate-resilient crop development.

The present result arises from just a small pilot field survey, which raises the question of what more might be found in Korea through surveys involving diverse growers with home gardens, experimental farms, small- or large-scale commercial production, and ornamental plant collections, over the entire Korean peninsula and offshore islands. Consulting the popular iNaturalist plant reporting platform (Internet, 15 January 2026) shows that actual distribution of cultivated taro in Korea is much wider than has been stated cumulatively in published reports seen by us. In Korean urban centers (Seoul and elsewhere), potted taro plants are commonly grown alongside other potted vegetables in very small-scale “house gardens” along the sides of buildings and roadsides (authors’ observations). We did not survey such potential sources. Taro is used in Korea not just as a starchy food plant, but also as a source of peeled petioles, fresh or dried. Investigating the varieties used for starch, as green vegetables, or both will also be of historical and practical interest, and can be linked to the food cultures involving taro across East Asia and the world.

The Korean taro samples exhibited pronounced morphological variability (Figs. 1–4), and different lineages may also differ in chemical composition and taste (Gerrano et al., 2021; Sánchez-Chino et al., 2021; Mitharwal et al., 2022; Tan et al., 2025). Clarifying the evolutionary relationships and geographic origins of these diverse forms requires reliable molecular markers, and Sanger sequencing of short, highly informative target sequences can be cost-effective for wide geographical surveys with many samples. To guide future studies, we identified several plastome regions with elevated nucleotide diversity that may serve as lineage-specific markers (Fig. 7). In particular, variation was concentrated in both LSC and SSC regions, with the trnN–ndhF intergenic spacer showing the highest variability, indicating its potential as a promising molecular marker.

Several of the highly variable loci identified in this study overlap with previously reported ptDNA markers for taro. Ahmed et al. (2013) described 30 ptDNA markers, five of which—trnS–trnG, accD–psaI, ndhF–rpl32, ccsA–ndhD, and rps15–ycf1—correspond to loci identified here as highly variable. These loci revealed major plastid lineages CI, CII, and CIII (Ahmed et al. 2020) that were subsequently confirmed with whole plastome analysis (Matthews et al., 2024). Two highly variable loci detected in our analysis (trnH–psbA and trnN–ndhF) were not included among previously-identified variable loci. Incorporating these loci—particularly trnN–ndhF, the most variable region identified—into taro genotyping frameworks could enhance the utility of plastid markers for distinguishing closely-related clonal lineages of taro, including those cultivated in Korea. The trnN–ndhF IGS is located near the IRb–SSC junction, where intraspecific variation can be influenced by IR boundary expansion/contraction and associated gene-conversion processes. Comparable junction-proximal intraspecific polymorphism has been reported in Medicago minima (L.) Bartal., in which small-scale shifts at the IR/SSC boundary were accompanied by gene conversion that homogenizes repeat copies and may overwrite adjacent single-copy sequences, thereby complicating strict positional homology in this region (Choi et al., 2020). For these reasons, variation in IR boundary–associated regions may have limited reliability for phylogenetic inference (Lockhart et al., 2006; Abdullah et al., 2020, 2021). Nevertheless, such highly variable plastome polymorphisms can remain informative for lineage or sub-lineage discrimination.

This study provides the first plastome-level assessment of genetic diversity in Korean taro and contributes to a clearer understanding of intraspecific phylogenetic structure within C. esculenta. Our results establish a foundation for more precise differentiation of taro cultivars in Korea, provide genomic resources relevant to the conservation and utilization of taro crop diversity, and clarify the position of Korea’s germplasm within the wider context of taro’s evolution, domestication and worldwide dispersal.