Pedigree Accumulation Analysis: Combining Methods from Community Ecology and Population Genetics for breeding adult estimation

Summary: Sard et al. describe an alternative to using number of successfully breeding adults (Ns) to predict recruitment because these data are difficult to collect. They used two non-parametric species richness estimators with genetic pedigree reconstruction data. They also tested these methods against two likely sources of variation associated with reproduction; reproductive success and sex ratio biases. These methods were used with empirical data for two species with many differences between the data sets.

Line 57: Is “RH: Pedigree Accumulation Analysis” a key word?

Line 75: Great examples of limitations to estimating spawning population size. Consider adding a sentence addressing limitations to estimating the number of offspring produced or emphasize how they are easier to sample than adults.

Line 133: Define and/or include reference for Type III survivorship functions.

Line 135: Using create twice in a row is a bit awkward. Consider, “We built an R package, breedr, which creates a model…”.

Line 140: Consider adding a little more information in the introduction about the study species and the systems they inhabit to better convey the conservation and management importance of these species and habitats. The text and table in the supplementary materials (S1) is great and some of this could be used in the introduction.

Line 188: Can further explanation be added for the range of the uniform distribution?

Line 253: Change “inferred” to “infer”.

Line 254: Additionally, sampling methods could impact accuracy if biased towards collection of more closely related individuals.

Line 277: Change “genotype” to “genotyped”.
Line 309: Change “an” to “a”.

Line 314: Add “of” and “was” to, “…variation in the distribution of fertilized eggs among mates when NS was large…”

Line 344: Consider rewording to, “In Figure 2, as Noff_S/NS approached 1, the jackknife estimates were biased…”.

Line 396: Add “be” after “can”.

Line 415: Change “Interesting” to “Interestingly”.

Supplemental Information:

Line 11: I am confused by the wording of “…which represent three orders of magnitude in which offspring are typically sampled and genotyped in the field.” Perhaps include a citation to support the typical sampling referenced and consider rewording because I assume genotyping is not happening in the field.

Line 24: Missing a word after “may”, perhaps “be” or “present”.

Line 60: Is the skewed male RS prediction and estimators based on behavior? Does it vary depending on the sample site and species?

Lines 74-76: It looks like the type A and B errors were estimated from studies with Chinook salmon, is there any concern that these estimates may not work as well for Lake Sturgeon or other species?

Figure S8: Consider adding a legend for the line colors.

Conclusions: I believe this method can be broadly used in this field. It was very interesting to see the differences between the Chao and Jackknife methods and to think about the different scenarios in which one may be more useful than the other. The application of these methods to simulations and varied empirical data made the possible uses and limitations for these methods clear.