Javan M. Bauder

• PhD Environmental Conservation
• Assistant Unit Leader at U.S. Geological Survey

Hello,

I am preparing to run an individual-based landscape genetics analysis on Eastern Indigo Snakes and was wondering if someone could provide some guidance on how to deal with individuals identified as full- or half-sibs. My study species is more-or-less continuously distributed across my landscape and I am measuring the genetic distance between samples using Bray-Curtis distances and principle components axes. I understand that researchers often randomly remove all but one member of full-sib families for population genetics analyses but I have seen less justification for this practice in the context of individual-based landscape genetics analyses.

I have tested for the presence of family structure using COLONY v1.2 and v2. They generally give similar results and indicate the presence of one 4-sib family, three 3-sib families, and 15-22 2-sib families (all full-sibs) out of 109 samples. I understand that these small family sizes may suggest that family structure is being poorly estimated and that I should be cautious about inferring family structure from these results. I have other biologically-based reasons to suspect some of these results as well (e.g., unrealistically large distances between some full-sibs).

Nevertheless, I am wondering if there is any utility in removing all but one individual from these full-sib families, even if the family structure is uncertain and if I am conducting an individual-based landscape genetics analysis. I noticed in a 2017 paper by Robin Waples and Eric Anderson in Molecular Ecology where they suggested caution when removing full-sibs in some circumstances, particularly with weak family structure.

Any suggestions would be welcome!

Thank you,

Javan Bauder

I have a question about the best way to test for violations of Hardy-Weinberg Equilibrium (HWE) among microsatellite loci for a species that is continuously distributed across a study area and showing IBD. We are looking at how landscape features affect gene flow among Eastern Indigo Snakes across a 25 x 50 km study area. We have 110 samples and about half are clustered in the southern half of the study area. A spatial correlogram of individual genetic distance shows that spatial autocorrelation among samples becomes non-significant at 5-10 km. We have used COLONY to identify full-sibs and found about 15 full-sib families although family size was usually two (max. four). There is significant IBD within our study area. STRUCTURE identifies K=4 with all 110 samples but when we randomly exclude all but one full-sib from each full-sib family STRUCTURE identifies K=1-2. We suspect that these STRUCTURE results are the result of neighborhood effects and IBD, respectively.

When we test for violation of HWE at our 15 loci, four have significant violations of HWE. Estimated null allele frequencies at these four loci are 6-15%. When we randomly excluded all but one member from each full-sib family, these four loci were still significantly out of HWE.

In a situation such as this, is it appropriate to test for HWE using all samples? I know that in systems with discrete populations researchers often test for HWE within each population, since violations may represent a mixture of multiple populations. But any designations of “populations” in our study area seem very arbitrary (e.g., driven by sampling intensity rather than the distribution of individuals).

Does anyone have any suggestions about the appropriate way(s) to test for HWE in a system such as ours?

Thanks,

Javan Bauder

I am creating resource selection functions to examine habitat selection using radio telemetry data and 1:1 paired logistic regression models. Habitat features at each telemetry point are paired with a single measure of the availability of those features that is unique to a particular telemetry point. I first fit models using R's glm function by differencing the used and available points and fitting a no-intercept glm:

glm(point ~ -1 + diffx1 + diffx2 + diffx3 ... diffxn, data, family=binomial)

I have also fit the same model using the survival package's clogit/coxph functions. In this case, each 1:1 pair of used/available points is grouped into a stratum:

clogit(point ~ x1 + x2 + x3 ... xn + strata(pair), data)

I have 16 covariates and some of these are correlated. However, when I calculate pair-wise correlation coefficients and variance inflation factors I get different results depending on whether I use the differenced data used to fit the glm or the original data used to fit the clogit/coxph. Both glm and clogit/coxph give me identical results but the different degrees of collinearity have a strong influence on the resulting analyses.

Does anyone have a recommendation of which form the data (differenced or original) I should use for assessing collinearity.

Many thanks!