grad_parent_reparam: The gradient of 'obj_parent_reparam' with respect to 's',...
In dcgerard/updogAlpha: Using Parental Data for Offspring Genotyping

Description Usage Arguments Value Author(s)

The gradient of obj_parent_reparam with respect to s, ell, and r, which recall are reparameterizations of the bias, seqeuncing error, and overdispersion, respectively.

1	grad_parent_reparam(pcounts, psize, ploidy, pgeno, s, ell, r, weight = 1)

`pcounts`	The number of counts of the reference allele we observe in a parent.
`psize`	The total number of reads we observe in a parent.
`ploidy`	An integer. The ploidy of the species. This is assumed to be the same for each individual.
`pgeno`	The genotype of the parent.
`s`	Same as `log(bias_val)` in `obj_offspring`.
`ell`	We have `seq_error = expit(ell)` from `obj_offspring`.
`r`	Same as `log((1.0 - od_param) / od_param)` from `obj_offspring`.
`weight`	A double that multiplies to the log-likelihood (objective). Should only be used when `outlier = FALSE` during the M step of the EM algorithm.