Computational protocol: Mouse Genome-Wide Association Mapping Needs Linkage Analysis to Avoid False-Positive Loci

Similar protocols

Protocol publication

[…] The association between spontaneous and urethane-induced lung tumor phenotypes (mean percentages or mean multiplicities) was expressed as a correlation coefficient. The association between each SNP and lung tumor phenotypes (expressed as log+1 of phenotype value) was tested by t-test. Haplotype analysis was carried out according to , using a sliding window approach (window size of 3 SNPs) and the BROAD dataset. The association between each haplotype and lung tumor phenotypes (expressed as log+1 of phenotype value) was tested by F-test. To control the genome-wide false-positive fraction (12,959 and 138,793 t-tests for databases WTCHG and BROAD), statistical thresholds were computed both in accordance with the Bonferroni principle and with a permutation test (20,000 permutations). In particular, the distribution of the 20,000 smallest p-values among the 12,959 (or 138,793) p-values under the null hypothesis was obtained. This approach implicitly uses the correlation structure of the data . The 5th and 10th centile of the reference distribution of the smallest p-values were used to guarantee a 0.05 or 0.10 overall false-positive fraction.Genome-wide genetic linkage was carried out by interval mapping using R/qtl . Marker order and position on chromosomes were established by multipoint analysis of the data using the MAPMAKER/EXP program . Genetic distances were computed using Haldane's mapping function. Single-locus genome scans were carried out using the ‘scanone’ function of R/qtl (http://www.biostat.jhsph.edu/˜kbroman/qtl/) using the Haley-Knott regression analysis. To increase the power to detect weak QTLs and to condition on the presence of the Pas1 genotype, a composite interval mapping was carried out using the three most significant markers identified by a stepwise regression as covariates . In addition, the Kras2 genotype was used as a covariate in a single-locus genome scan of (A/J×C3H/He)F2 and (BALB/c×C3H/He)F2 intercross populations. Genome-wide significance thresholds (α = 0.05) were generated through permutation tests (10,000 permutations) as described . […]

Pipeline specifications

Software tools R/qtl, MAPMAKER
Application WGS analysis
Organisms Mus musculus
Diseases Adenoma, Lung Neoplasms, Neoplasms
Chemicals Urethane