Despite significant advances in phylogenetics over the past decades, the deep relationships within Bivalvia (phylum Mollusca) remain inconclusive. Previous efforts based on morphology or several genes have failed to resolve many key nodes in the phylogeny of Bivalvia. Advances have been made recently using transcriptome data, but the phylogenetic relationships within Bivalvia historically lacked consensus, especially within Pteriomorphia and Imparidentia. Here, we inferred the relationships of key lineages within Bivalvia using matrices generated from specifically designed ultraconserved elements (UCEs) with 16 available genomic resources and 85 newly sequenced specimens from 55 families. Our new probes (Bivalve UCE 2k v.1) for target sequencing captured an average of 849 UCEs with 1085-bp in mean length from in vitro experiments. Our results introduced novel schemes from six major clades (Protobranchina, Pteriomorphia, Palaeoheterodonta, Archiheterodonta, Anomalodesmata and Imparidentia), though some inner nodes were poorly resolved, such as paraphyletic Heterodonta in some topologies potentially due to insufficient taxon sampling. The resolution increased when analyzing specific matrices for Pteriomorphia and Imparidentia. We recovered three Pteriomorphia topologies different from previously published trees, with the strongest support for ((Ostreida + (Arcida + Mytilida)) + (Pectinida + (Limida + Pectinida))). Limida were nested within Pectinida, warranting further studies. For Imparidentia, our results strongly supported the new hypothesis of (Galeommatida + (Adapedonta + Cardiida)), while the possible non-monophyly of Lucinida was inferred but poorly supported. Overall, our results provide important insights into the phylogeny of Bivalvia and show that target enrichment sequencing of UCEs can be broadly applied to study both deep and shallow phylogenetic relationships.

We designed a new set of probes, Bivalve UCE 2k v.1, targeting ~2000 loci for bivalve species. Based on this probe set, we collected UCEs of 85 specimens from their target-captured Illumina sequencing data. The data matrices of this study were generated according to the probes (v1, v2, or all probes) and taxon data occupancy rate (25%, 35%, 50%, 60%, 75%) using Phyluce v.1.7.1. These data matrices were processed to conduct phylogenetic analyses. The methods were detailed in the main text and supplementary documents.

Funding provided by: University Grants CommitteeROR ID: https://ror.org/00djwmt25Award Number: C2013-22GF Funding provided by: University Grants CommitteeROR ID: https://ror.org/00djwmt25Award Number: 12102222