In particular, the selection was focused on variants showing low MAF (i.e. rare variants), appearing in interesting regions (exons and transcription factor binding sites) or showing a large difference in the allele count between cases and controls.
The DNA of each DH line was diluted to 100 ngμL−1 and subjected to genotypic analysis using the DArT-seq platform essentially as described by Curtois et al. [ 18], except that the marker curation involved removing the markers with low minor allelic frequency (MAF) (i.e. < 0.1) and > 50% missing data.
Similarly, Figure D1 of supplementary Additional file 4 shows the growth times between MDR and FMDR in other MAFs (i.e., MAF = 0.2 and MAF = 0.4).
Importantly, the weakening of statistical power as a result of incomplete genome-wide marker coverage and small sample size will be intensified if causal variants have lower MAFs (i.e., rare causative alleles) than the genotyped SNPs used for the GWAS (Yang et al. 2010)—an unfavorable situation that likely occurred in our GWAS of tocochromanol grain traits.
Experimental studies performed on data collected from the French RATP metro network show that the AR-MAFI methods outperform (in terms of readability and accuracy) the standard distance domain or wavelength domain methods in localizing and characterizing corrugation.
In each simulation, 100 or 1000 QTL were sampled assuming that they followed one of the three QTL MAF patterns i.e. moderately low average MAF, very low average MAF, or extremely low average MAF.
Only SNPs with minor allele frequency (MAF) ≥0.10 (i.e., minor allele present in at least 87 accessions) were considered for GWAS.
The mixture MAF for SNP i is ∑ k = 1 K f k MAF k i, where f k is the mixture frequency assigned to population k and MAF ki is the MAF of SNP i in k.
For both concurrent estimation and pre-adjusted data, estimates of α from pooled data in Brahman with a MAF > 0.01 (i.e., including SNP with a MAF < 0.05) have a very low correlation with estimates from individual data.
From these bins, we preferentially selected tag SNPs using several criteria: Illumina design scores (quantifying how well a SNP can be genotyped), validation status (indicating how many platforms validate a SNP), minor allele frequency (MAF), and location (i.e. coding region).
Then, if we write pi for the MAF of variant i, the probability for a case to be homozygous at this variant is pi, and the probability to be homozygous for one or more variants is eHOM = 1 − Π (1 − pi).
