Major gaps remain in our understanding of the ecology, evolution, biodiversity, biogeography, extinction risk, and adaptive potential of reef building corals. One of the central challenges remains that there are few informative genetic markers for studying boundaries between species, and variation within species. Reduced representation sequencing approaches, such as RADseq (Restriction site Associated DNA sequencing) have great potential for resolving such relationships. However, it is necessary to identify loci in order to make inferences for endosymbiotic organisms such as corals. Here, we examined twenty-one coral holobiont ezRAD libraries from Hawai?i, focusing on P. lobata and P. compressa, two species with contrasting morphology and habitat preference that previous studies have not resolved.</p><p>We used a combination of de novo assembly and reference mapping approaches to identify and compare loci: we used reference mapping to extract and compare nearly complete mitochondrial genomes, ribosomal arrays, and histone genes. We used de novo clustering and phylogenomic methods to compare the complete holobiont data set with coral and symbiont subsets that map to transcriptomic data. In addition, we used reference assemblies to examine genetic structure from SNPs (Single Nucleotide Polymorphisms). All approaches resolved outgroup taxa but failed to resolve P. lobata and P. compressa as distinct, with mito-nuclear discordance and shared mitochondrial haplotypes within the species complex. The holobiont and ‘coral transcriptomic’ datasets were highly concordant, revealing stronger genetic structure between sites than between coral morphospecies. These results suggest that either branching morphology is a polymorphic trait, or that these species frequently hybridize. This study provides examples of several approaches to acquire, identify, and compare loci across metagenomic samples such as the coral holobiont while providing insights into the nature of coral variability.