Linkage mapping of 1454 new maize candidate gene Loci

Abstract

Bioinformatic analyses of maize EST sequences have highlighted large numbers of candidate genes putatively involved in agriculturally important traits. To contribute to ongoing efforts toward mapping of these genes, we used two populations of intermated recombinant inbred lines (IRILs), which allow a higher map resolution than nonintermated RILs. The first panel (IBM), derived from B73 x Mo17, is publicly available from the Maize Genetics Cooperation Stock Center. The second panel (LHRF) was developed from F2 x F252 to map loci monomorphic on IBM. We built framework maps of 237 loci from the IBM panel and 271 loci from the LHRF panel. Both maps were used to place 1454 loci (1056 on map IBM_Gnp2004 and 398 on map LHRF_Gnp2004) that corresponded to 954 cDNA probes previously unmapped. RFLP was mostly used, but PCR-based methods were also performed for some cDNAs to map SNPs. Unlike in usual IRIL-based maps published so far, corrected meiotic centimorgan distances were calculated, taking into account the number of intermating generations undergone by the IRILs. The corrected sizes of our framework maps were 1825 cM for IBM_Gnp2004 and 1862 cM for LHRF_Gnp2004. All loci mapped on LHRF_Gnp2004 were also projected on a consensus map IBMconsensus_Gnp2004. cDNA loci formed clusters near the centromeres except for chromosomes 1 and 8.

Figure 1.—

Chromosome 1 of framework IBM_Gnp2004 and LHRF_Gnp2004 maps used for cDNA mapping. The complete figure with all 10 chromosomes is available as supplementary Figure S1 at http://www.genetics.org/supplemental/. Links between maps indicate common markers. Core bin markers are boxed. Hatched segments in the chromosome axis indicate segments within which the LOD value indicating the stability of the order between loci is <1. Map coordinates at the left side of the axis are in overestimated centimorgans calculated by MapMaker using the “RI self Haldane” parameters, whereas positions in italics at the right side of the axis are corrected meiotic centimorgan distance (CCD) coordinates, taking into account the four generations of intermating undergone by the mapping population (see materials and methods). Side plots indicate the frequencies of the B73 or F2 allele in the segregation data (dots) for all loci mapped against these frameworks, as well as the 1% confidence interval (lines) of the expected frequency (0.5), taking into account sample size and genetic drift during the four generations of intermating (see materials and methods). Open circles indicate the position of the centromeres (from MaizeGDB).

Figure 2.—

Effect of simulated data errors on map size expansion. The vertical axis indicates the percentage of map size expansion consecutive to increasing frequencies of randomly distributed simulated errors in the segregation data of the IBM_Gnp2004 framework map. The solid line indicates the mean value, and dashed lines show the minimum and maximum values obtained during 1000 simulations.

Figure 3.—

Relationship between MapMaker-computed distance using the “RI self Haldane” option (solid line) and corrected centimorgan distance (CCD) for IRILs having undergone four generations of random intermating. The right vertical axis and dashed line indicate the rate of distance overestimation resulting from the use of MapMaker with the RI self Haldane option to compute distances from IRIL populations.

Figure 4.—

Density of cDNA mapped along the linkage groups of the IBM_Gnp2004 map. Loci projected from LHRF_Gnp2004 are included. The horizontal axis indicates the number of cDNA mapped. The vertical axis indicates map coordinates in centimorgans on the IBM_Gnp2004 map. Arrows indicate the position of the centromeres (according to MaizeGDB).