Figure 2

The bioinformatics pipeline RFAPtools. A. Assembly of pseudo-reference sequence. Pair-end sequence reads from ZY821, No2127 and pooled DHs were linked to form artificial sequence tags and collapsed into clusters. Then clusters present in both ZY821 and DHs were concatenated to assemble the ZY821 pseudo-reference sequence (ZY821-PRF). Thirdly, all clusters from No2127 were aligned to ZY821-PRF using SOAP with two nucleotide mismatches and those clusters that could not mapped were joined to the end of ZY821-PRF as clusters specific to No2127 to form the final pseudo-reference sequence. B. Discovery and genotyping of SNPs and PAVs. Sequencing reads of each individual were aligned to the PRF and all alleles at each locus were called as “+” indicating the presence of an allele and “-” indicating the absence of that allele in the two parents and 91 DH lines. C. Discrimination of allelic SNPs from homoeologous loci. A co-occurrence matrix of every two alleles from a complex locus with homoeologous sequences in the BnaNZDH population was made and allelic SNP and PAV loci were discriminated from homoeologous loci. Then genotypes of all individuals were scored for each locus.