**1. Introduction**

Maize (*Zea mays* L.) is one of the most important cereal crops consumed in sub-Saharan Africa (SSA) and an essential component of livestock feed in the developed as well as developing world. Approximately 10,000 years ago in Central Mexico, maize was domesticated from its ancestor, the

wild grass-teosinte (*Zea mays* spp. *parviglumis*) [1]. Post-domestication of teosinte led to the transfer of beneficial adaptive genomic regions into common maize [2]. Mutations together with recombination events, either due to natural or farmer-mediated selection, then generated novel allele combinations [3,4]. Thus arose the "landraces", traditional varieties selected by farmers for adaptation to local conditions and food preferences [5]. The landraces, though typically low in yield, are invaluable sources of diversity that could be drawn upon to broaden the genetic base of elite maize germplasm, and to further enhance adaptation to changing environments and pathogens [6]. Indeed, earlier breeders identified and composited the most productive landraces into genetically diverse populations, forming the foundation of inbred line development and pedigree breeding [7]. Modern breeding practices, in which a narrow range of inbred lines are included in crossing programs, have narrowed the genetic base of most cultivated crops, which have negatively a ffected the adaptability of the crop to changing climates, devastating pathogens, and insect-pests [8]. In order to provide a bu ffer against the possible e ffects of novel threats, it is essential to broaden the genetic base of breeding populations by introgressing an enlarged pool of beneficial alleles.

The high selection pressure under low input and climatically stressed maize growing environments in Africa is likely to have resulted in local adaptations with potential value for breeding stress tolerant and nutritionally enriched varieties [9]. Indeed, there are examples of successful use of local African genetic resources of maize in the development of varieties such as Katumani in Kenya and Longe-5 from Uganda [9]. However, while the New World maize gene pool is well represented in genebanks [10] and well characterized [3], there are collections of African maize germplasm without adequate data of their genetic make-up [9]. This missing genetic data makes searching for promising landraces within the African maize genebank collections like "finding the proverbial needle in a haystack" [11]. Therefore, to develop improved climate resilient maize varieties for SSA, e fficient characterization, identification, and utilization of climatically adapted local African maize germplasm is a crucial prerequisite [9].

Recently, we phenotypically characterized a panel of maize landraces originating from Burkina Faso, Ghana, and Togo, which cover a wide range of climatic conditions classified as Sahel and coastal West Africa (WA) [12]. The study revealed that the maize panel varied considerably in flowering date, plant architecture, yield and yield related traits, and other characteristics. These di fferences allowed the formation of five distinctive morphological groups [12]. The Sahel gene pool was highly distinct and was considered a valuable resource for future genetic enhancement. However, phenotypic variation can be confounded by the environment and a high degree of plasticity [13]. Given that gene flow from multiple introductions may have shaped the population structure of African maize, diversity analysis of this panel using state-of-the-art genotyping techniques could help provide deep insights into the complexity of its genetic architecture and composition.

The assessment of genetic diversity by genotyping-by-sequencing (GBS) provides robust estimates of diversity and has been increasingly adopted as a fast, high-throughput, and a ffordable tool for whole-genome genetic diversity analysis in large germplasm sets [14]. The diversity array technology sequence (DArTseq) markers, characterized by high marker coverage, call rates, and scoring reproducibility, has emerged as a useful GBS approach for assessing genetic diversity and population structure in various crops including wheat [15,16], rice [17], watermelon [18], common bean [19], and maize [20,21]. The objective of this study was to examine the genetic diversity and population structure of a maize panel comprising landraces from Sahel and coastal Africa together with a reference population using the GBS-DArTseq approach. The relevance of our results for further exploration and utilization of local genetic resources of maize in Africa is discussed in this manuscript.
