Fig. 7

Comparisons of association mapping using linear mixed model based on reference and dispensable genomes. a Manhattan plots for association mapping of rice grain width using CDPs based on the dispensable genomes. The horizontal dashed line indicates the genome-wide significance threshold (P = 6.6 × 10⁻⁸). The lead CDP is marked in brown and labeled. b Alignment of both the raw assembly and the local reassembly of the contig harboring GW5/qSW5 to the Nipponbare genome. The aligned regions of the contigs and the Nipponbare genome are indicated by red bars and connected by dashed lines. OsIPC05060021H3C4 and OsIPC05060021H1C3 are the local assembly of the contig OsIPC05060021. The black parts of these two contigs indicate the regions which could be aligned to the Nipponbare genome while the green and red parts denote insertion regions in the Nipponbare genome. The arrows indicate the strand of these sequences. The red segment of qSW5 indicates the position of the coding region of qSW5 in these sequences. c Comparisons of LMM P values of lead SNPs with that of lead CDPs for 1169 metabolic traits of which the P values for the lead SNPs or the lead CDPs passed the significance threshold. The red dashed line indicates where the P values of the two approaches are equal. d The distribution of chromosome distances between significant lead SNPs and the closest significant (P ≤ 6.6 × 10⁻⁸) or similarly significant (the difference of P values less than hundred times) CDPs