Dataset: Microsatellite genotypes

At our field sites in the shallow seagrass habitats near the Florida State University Coastal and Marine Laboratory (FSUCML) in St. Teresa, Florida, USA (29° 54' N, 84° 30' W), multiple B. neritina colonies are commonly found attached to the same blade of seagrass, often contacting each other at times of the year when densities are highest. In December 2016, we randomly selected 15 blades of seagrass containing adult colonies within a roughly 10-50 meter area immediately east of the FSUCML. Sampled seagrass blades were roughly 2 – 5 meters from each other. Each sampled blade contained between 2 to 11 adult colonies (median = 4), which were all genotyped at 16 microsatellite loci, following methods described in Burgess et al. (2019). We collected and genotyped an additional 53 randomly selected colonies from the same general area to provide a more precise estimate of the population allele frequencies used to estimate relatedness. In total, there were 127 individuals in the sample, of which 74 were the focal individuals from the 15 blades of seagrass. Our focal populations contain the S1 haplotype of the cytochrome oxidase c subunit I (COI) gene, and do not include cryptic species.

We used the program COLONY (Version 2.0.6.6., Build 20200830) to identify pairs of individuals with a full- or half-sib relationship using a sibship reconstruction analysis. Sibship reconstruction requires no a-priori knowledge of the relationships in the sample of individuals, and uses a clustering algorithm to arrange individuals into families based on Mendelian rules of allele inheritance. Given the time of year and the size of colonies that were collected, the sample was considered to consist of a single cohort (i.e., no parent-offspring relationships in the sample). We used the full sample of 127 individuals to increase the confidence in identifying family clusters, but focus on the 74 individuals sampled from known seagrass blades.

For all runs in COLONY, the parameter settings used were: allele frequency unknown and not updated, monoecious, no inbreeding, diploid, male and female polygamy, no clones, full sibship scaling, weak sibship prior of 1, long length of run, full-likelihood with high precision, allelic drop out and false allele rate of 0.0001. We carried out 5 runs using different seed numbers to check the reliability of the results. Only full- and half-sibs that were equal to or above a certain probability (0.8, 0.9, or 0.95) in all 5 replicate runs were considered to be reliably estimated.