QC for LogNormalDFE (HomSap)

[RyanGutenkunst]

PR for new model: PR
Original paper: paper
Parameter values: DFE parameters are based off the Full model in Table S3, using recent mutation rate estimates
Population indices: N/A

Potential issues that might lead to differences between the model implementations:
SLiM and dadi use different conventions for selection coefficients. In SLiM, the homozygote has fitness 1+s; in dadi it is 1+2s. Since this model was inferred using dadi, to map it to SLiM we need to double the selection coefficient, by shifting the center of the lognormal distribution.

QC'er requests: @petrelharp