Analysis of RFLP mapping inacuracy in Brassica napus L.

We identified sources of mapping inaccuracy during the construction of RFLP linkage maps from one F2 population and two F1 microspore-derived populations from the same cross of oilseed Brassica napus. The genetic maps were compared using a total of 145 RFLP marker loci including 82 loci common to all three populations. In the process, we identified a series of mapping events that could lead to ambiguous conclusions. Superimposed restriction fragments could be mistaken as a single dominant restriction fragment in a F2 population and, when analyzed as such, would yield inaccurate linkage information. Residual heterozygosity in parental lines resulted in complicated allelic assignment and yielded subsequent difficulties in linkage determination. Loose and spurious linkages occurred during mapping and were identified by comparing maps derived from different populations. LOD scores and Chi-square test of independence were compared for their capacity to detect loose linkages or generate spurious ones. Extreme segregation distortions towards the same parental allele also contributed to an additional source of spurious linkage. Small but significant segregation distortions resulted in reduced estimates of the recombination fraction. The use of the same 'probe x enzyme' combinations in doubled haploid populations allowed the identification of the correct allele assignment as well as loose and spurious linkages. A translocation between two homoeologous linkage groups was observed. The consequences of such a chromosomal
event as a source of error in mapping applications are discussed.