I am doing association mapping with SSR markers. I have 106 markers when I combined them with phenotypic data I got less QTLs than in the case if I applied minor allele frequency (markers showed less 10% allele frequency, about 41 were excluded).
I used FDR to show the significance of marker-trait association).
Thanks in advance.
The FDR (as the Bonferroni, as Holm etc) is just a multitesting correction to adjust the single association p-values for your markers based on the number of markers (or in general of single analyses) operated (increasing the number will increase the significance threshold). Applying the MAF filter in order to extract only alleles with at lest a percentage (e.g. 10% as you specified) of representation in the analyzed population is necessary to avoid your p-values to be false positive: allelic association analysis based on their distribution may detect a significant prevalence of an allele in case or controls but this may be just due to the very low percentage of subjects with that allele. To avoid that, MAF filters are applied in association analyses.
So, if your question is in what relationship they are you can think that applying an higher MAF threshold would reduce the number of markers in the analysis and so decreasing the significance FDR value (making easier to obtain a significant result) and viceversa.
Hope it helps!
this link will give you a softeware for calculation of FDR
http://genomics.princeton.edu/storeylab/qvalue/
to my knowledge, just by chance a minor allele can be within a high LD with allele of other locus which leads to a false (spurious) Association between them that is most likely to show a fake qtl.
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping