grad_offspring_weights: Gradient of 'obj_offspring_weights_reparam'
In dcgerard/updogAlpha: Using Parental Data for Offspring Genotyping

Gradient of obj_offspring_weights_reparam

1	grad_offspring_weights(ocounts, osize, weight_vec, ploidy, prob_geno, s, ell, r)

`ocounts`	The observed counts of the refernce allele for each individual.
`osize`	The observed number of reads for each individuals.
`weight_vec`	A vector of weights between 0 and 1 (do not need to add up to 1).
`ploidy`	An integer. The ploidy of the species. This is assumed to be the same for each individual.
`prob_geno`	The allele frequencies of the genotypes. See `get_prob_geno`.
`s`	We have `s = exp(d)`, where `exp(d)` is the bias term.
`ell`	The logit of the sequencing error rate
`r`	We have `r = log((1 - tau) / tau)`