QTLs at markers with high entropy / variance

[Lena Flörl]

Dear Forum, Dear Karl,

I get significant QTLs at markers that show rather high entropy (max. 0.48) and variance (max 0.76) when checking for missing genotype information and I wonder whether that is actual spurious linkage or could be just an actually highly genetically diverse region? 

I perform QTL mapping of microbiome data (already spurious!) on an F1 population of grapevines which are genetically also highly heterozygous. Therefore I use the crosstype='4way'.

Now I wonder if there are any other ways of checking whether I can "trust" these QTLs? 

Thanks so much! 

Kind regards, 

Lena 

[Lena Flörl]

Maybe I should also mention that I checked for Segregation Distortion but none of these markers with a higher variance/entropy showed such distortion - which makes me believe that it's rather actual genetic variance than a genotyping error. But maybe that's a misunderstanding on my side. 

[Karl Broman]

In regions with low entropy and low variance, you have little genotype information and so low power to detect a QTL.

And so if observed QTL are in regions with more genotype information, that doesn't mean that they are spurious.

However, varying genotype information could bias estimates of QTL location.

Main ways to strengthen trust in QTLs are careful quality control on genotype and phenotype data, and plots of genotype vs phenotype at inferred QTL.

karl