Incorporate use of penetrance to weight covariates of unaffected

[CNuge]

See here for calculating penetrance: https://github.com/SequenceBio/seqbio-pedigree-ranking

We can maybe use this information to determine the weight of unaffecteds when calculating scores. Essentially, more penetrance == give more weight to unaffecteds, low penetrance == more uncertainty and therefore want to give less weight in this case.

[CNuge]

The higher the penetrance estimate, the higher we want the u weighting to be. The lower the penetrance estimate, the lower u weighting should be. Why? when penetrance is low U correct/incorrect mean less. When a variant is incompletely penetrant then the expected decoupling of variant status for unaffected individuals is higher.

So, maybe the best thing to do is to note this and to then leave it up to the user to select a value of weighting for u. Half a good starting point, go higher with a high penetrance estimate and lower with a low penetrance estimate?

[CNuge]

Currently the default weighting for U evidence is set to 0.5. Calculating penetrances for a pedigree gives an output value, that a user could then pass to the calculation in lieu of the default.