Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2007 Oct 17;2(10):e1036.
doi: 10.1371/journal.pone.0001036.

A Simple Method for Combining Genetic Mapping Data From Multiple Crosses and Experimental Designs

Free PMC article

A Simple Method for Combining Genetic Mapping Data From Multiple Crosses and Experimental Designs

Jeremy L Peirce et al. PLoS One. .
Free PMC article


Background: Over the past decade many linkage studies have defined chromosomal intervals containing polymorphisms that modulate a variety of traits. Many phenotypes are now associated with enough mapping data that meta-analysis could help refine locations of known QTLs and detect many novel QTLs.

Methodology/principal findings: We describe a simple approach to combining QTL mapping results for multiple studies and demonstrate its utility using two hippocampus weight loci. Using data taken from two populations, a recombinant inbred strain set and an advanced intercross population we demonstrate considerable improvements in significance and resolution for both loci. 1-LOD support intervals were improved 51% for Hipp1a and 37% for Hipp9a. We first generate locus-wise permuted P-values for association with the phenotype from multiple maps, which can be done using a permutation method appropriate to each population. These results are then assigned to defined physical positions by interpolation between markers with known physical and genetic positions. We then use Fisher's combination test to combine position-by-position probabilities among experiments. Finally, we calculate genome-wide combined P-values by generating locus-specific P-values for each permuted map for each experiment. These permuted maps are then sampled with replacement and combined. The distribution of best locus-specific P-values for each combined map is the null distribution of genome-wide adjusted P-values.

Conclusions/significance: Our approach is applicable to a wide variety of segregating and non-segregating mapping populations, facilitates rapid refinement of physical QTL position, is complementary to other QTL fine mapping methods, and provides an appropriate genome-wide criterion of significance for combined mapping results.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.


Figure 1
Figure 1. The need for locus-specific P-values.
The 95% LOD score (the LOD score equivalent to a locus-specific P = 0.05) was calculated using 10,000 permutations for markers on Chr. 1 for body weight in several different populations. Each marker is indicated by a dot with connecting lines interpolated between adjacent markers. TJL BXD are BXD strains available from The Jackson Laboratory (The BXD strains developed by Taylor and colleagues [26], [27]). New BXD are the recently developed BXD strains currently resident at UTHSC. Note that the maximum and minimum values of the 95th percentile LOD score vary considerably for the AIL population, somewhat for the RI (New BXD and TJL BXD) populations, (predicted by missing data pattern) and very little for the 183 member F2 population tested. (There are only three widely spaced markers genotyped for the F2 population on Chr. 1, so the interpolation between points should not be interpreted as a meaningful line. However, markers on all chromosomes were very similar, between a 95% LOD of 1.2 and 1.4.)
Figure 2
Figure 2. Combined mapping for Hipp1a and Hipp9a.
This figure shows mapping data for the hippocampus weight loci Hipp1a and Hipp9a using 34 BXD strains (BXD; shaded line) and 679 advanced intercross animals (AIL, thin solid line) as well as the composite map using the described method (thick solid line). The genome-wide adjusted composite P = 0.05 threshold is −log P = 3.5 (dark solid horizontal line). Since 5000 permutations were used for each data set, the maximum −log P<3.7 (graphed as −log P = 3.7 for convenience) for each individual data set, so increasing the number of permutations might increase the peak combined value and slightly improve the range of the combined interval. Bars underneath the peaks are labeled AIL, BXD, and combined to indicate the l-LOD support interval of these mapping populations.

Similar articles

See all similar articles

Cited by 12 articles

See all "Cited by" articles


    1. Darvasi A, Soller M. Advanced intercross lines, an experimental population for fine genetic mapping. Genetics. 1995;141:1199–1207. - PMC - PubMed
    1. Zou F, Gelfond JA, Airey DC, Lu L, Manly KF, et al. Quantitative trait locus analysis using recombinant inbred intercrosses (RIX): theoretical and empirical considerations. Genetics 2005 - PMC - PubMed
    1. Iakoubova OA, Olsson CL, Dains KM, Ross DA, Andalibi A, et al. Genome-tagged mice (GTM): two sets of genome-wide congenic strains. Genomics. 2001;74:89–104. - PubMed
    1. Bolivar VJ, Cook MN, Flaherty L. Mapping of quantitative trait loci with knockout/congenic strains. Genome Res. 2001;11:1549–1552. - PMC - PubMed
    1. Darvasi A. Experimental strategies for the genetic dissection of complex traits in animal models. Nat Genet. 1998;18:19–24. - PubMed

Publication types