A Draft Genome Sequence for the Cashew Tree (Anacardium occidentale) Based on High Coverage Pacbio Sequencing

The cashew tree (Anacardium occidentale) is widely cultivated in the tropical world for its nuts and apple, both rich in bioactive compounds and sources of biofuel. Large germplasm collections and extensive progeny trials are supporting a dynamic breeding program at EMBRAPA in Brazil. Yet varietal development is a lengthy process.

With a two-fold objective of accelerating breeding and empowering germplasm assessment, we have generated a draft assembly of the diploid ~500 Mbp genome of the BRS226 dwarf variety using exclusively long read sequencing. Over 4M of filtered SMRT reads (median:6,588bp) were generated using P6C4 chemistry, totaling >33 Gbp of data (~70x coverage). Hierarchical nonhybrid assembly based on PBcR-MHAP was used to correct reads, and the longest 25× of corrected reads (12Gb in 934,094 reads) were assembled using Celera Assembler and polished with Quiver. The current assembly contains 532 Mbp in 8,050 contigs, the longest contig with 3.2 Mbp, while 7,809 contigs are >10 Kbp and 1,066 >100 Kbp. N50 size of the contigs is 160Kb. Quality Value (QV) was estimated from 10xMiSeq (2x250bp) filtered reads. PE reads were transformed into 'super-reads' with MaSuRCA and aligned to the reference assembly using Nucmer, validated using the Gage scripts in Quast. By counting both SNP and Indels we estimate that the assembly achieved a lower bound accuracy of 99.85% (QV28).

A high-density linkage map is under development and 24 cashew varieties including founders of the breeding program were resequenced at ~15X (Hiseq 2x150bp) to provide the baseline data for a high-density SNP chip.