tavareshugo / magichelpr Goto Github PK

A LOD score can, in principle, be calculated from the log-likelihood of the null and alternative models, as:

LOD = log10(likelihood_qtl / likelihood_no_qtl)

From the log-likelihood of the model, it should be fine to calculate as:

logLik(qtl_model)/log(10) - logLik(null_model)/log(10)

Need to update the vignette with the way the estimateFounderEffects() function now works (returning standardized effect sizes). Should also improve the documentation on specifying custom models for qtlScan()

When using this function with the option summarised = TRUE, there could be a column containing the SNP genotypes, so that the user does not have to run the inferMagicFounder function separately.

Hi Hugo,

I have issues installing the dependencies to MagicHelpR.

I know you are not maintaining the happy package, but i wanted to ask if you could tell me where to get the happy.hbrem package, since the link to http://www.well.ox.ac.uk/happy is broken.

Could you provide a functional package or tell me who to contact to get it for R?

Thank you very much.
Michel

Add a new function that performs a QTL scan using all pairwise combinations of markers. Basically, for each marker, use each of the other markers as a covariate. The output could then be visualised with a matrix plot (colouring cells by logP for example). Function name could be something like qtlPairwiseScan()