Low-cost assembly of a cacao crop genome is able to resolve complex heterozygous bubbles

Abstract

Cacao (Theobroma cacao) is a tropical tree that produces the essential raw material for chocolate. Because yields have been stagnant, land use has expanded to provide for increasing chocolate demand. Assembled genomes of key parents could modernize breeding programs in the remote and under-resourced locations where cacao is grown. The MinION, a long read sequencer that runs off of a laptop computer, has the potential to facilitate the assembly of the complex genomes of high-yielding F₁ hybrids. Here, we validate the MinION's application to heterozygous crops by creating a de novo genome assembly of a key parent in breeding programs, the clone Pound 7. Our MinION-only assembly was 20% larger than the latest released cacao genome, with 10-fold greater contiguity, and the resolution of complex heterozygosity and repetitive elements. Polishing with Illumina short reads brought the predicted completeness of our assembly to similar levels to the previously released cacao genome assemblies. In contrast to previous cacao genome projects, our assembly required only a small scientific team and limited reagents. Our sequencing and assembly methods could easily be adopted by under-resourced breeding programs, speeding crop improvement in the developing world.

Fig. 1. Pound 7 is better representative of both wild cacao and elite F₁ hybrids than the previous cacao reference genomes.

a Relative change in worldwide cacao yields, compared with the change in total production and the amount of land used to produce the cacao, normalized to the 1961 data points. Data from FAO.org. b Relative worldwide yield gains for cacao compared with the two other crops. Data from FAO.org. Values are relative, normalized to the 1961 datapoint. c The site of collection of Pound 7 is represented by the red square. The center of cacao’s diversity is the blue circle^,. The previous cacao reference genomes are Matina 1–6 (associated with Costa Rica) and B97–61/B2 (associated with Belize)^,. The satellite image (adapted from Google Maps) shows the coordinates of where Pound 7 was collected, 73.10 W 3.45 S (Pound, 1943; Turnbull, C.J. and Hadley, International Cocoa Germplasm Database). d Pound 7 is a heterozygous wild hybrid of several ancestral groups (Motamayor et al.), in contrast to the two domesticated cacao that were used for the previous reference genomes. e The number of heterozygous SNP calls out of 135,696 SNPs from resequenced cacao genomes shows the range of heterozygosity in cacao accessions. The current reference genomes are the “highly homozygous” B97–61/B2^, and Matina 1–6. Both are relatively homozygous compared with Pound 7 and the widely cultivated F₁ hybrid CCN 51

Fig. 2. Our assembly resolved a highly heterozygous locus.

The top panel is the alignment of BAC sequences for heterozygous haplotypes on Chromosome 4 of Pound 7. When aligned to the Matina 1–6 genome, Haplotype A maps to Scaffold 4: 142672 to 324489, and Haplotype B maps to Scaffold 4: 181600 to 392073. The lower two panels show the two haplotypes resolved in the MinION assembly, aligned to the BAC sequences. A summary of the annotations is in Supplementary Table 10. Haplotype A is deposited on NCBI as P7SI_AltHap_V3 and Haplotype B is P7SI_MatHap_V3