Solanum pennellii is a wild tomato species endemic to Andean regions in South America, where it has evolved to thrive in arid habitats. Because of its extreme stress tolerance and unusual morphology, it is an important donor of germplasm for the cultivated tomato Solanum lycopersicum 1 . Introgression lines (ILs) in which large genomic regions of S. lycopersicum are replaced with the corresponding segments from S. pennellii can show remarkably superior agronomic performance 2 . Here we describe a high-quality genome assembly of the parents of the IL population. By anchoring the S. pennellii genome to the genetic map, we define candidate genes for stress tolerance and provide evidence that transposable elements had a role in the evolution of these traits. Our work paves a path toward further tomato improvement and for deciphering the mechanisms underlying the myriad other agronomic traits that can be improved with S. pennellii germplasm.Crosses between distantly related plants can lead to substantial improvements in performance. Notably, S. pennellii × S. lycopersicum ILs have been used to define numerous quantitative trait loci (QTLs) for superior yield, chemical composition, morphology, abiotic stress tolerance and extreme heterosis 3,4 . Although genetic studies have proven informative, few genes underlying specific QTLs have been cloned, largely because of the lack of a S. pennellii genome sequence. To support QTL analyses, we sequenced the genome of S. pennellii using Illumina sequencing with ~190-fold coverage ( Fig. 1 and Supplementary Tables 1-5). The initial assembly size was 942 Mb, with a scaffold N50 value of 1.7 Mb and N90 value of 0.43 Mb (Table 1 and Supplementary Tables 6 and 7). We estimated the total genome size to be about 1.2 Gb using a k-mer-based analysis ( Supplementary Fig. 1 and Supplementary Table 8), in accordance with previous estimations 3,4 . We anchored 97.1% of the genome assembly to chromosomes using genetic maps and restriction site-associated DNA sequencing (RAD-seq)-based markers from the IL population 5 (Supplementary Note). Comparison of the assembly to publicly available BAC sequences indicated an accuracy of >99.9%, and a satisfactory accuracy of gap-filled regions was shown by realigning reads (Supplementary Fig. 2 and Supplementary Table 9). Of the 307,350 S. lycopersicum and 7,812 S. pennellii publicly available ESTs, 93% and >96% could be aligned to the genome, respectively (Supplementary Table 10), indicating comprehensive coverage of the gene-rich regions. We predicted 32,273 high-confidence genes and a potential set of 44,966 protein-coding genes and checked these