We applied whole genome shotgun sequencing to museum specimens of longhorn beetles and reconstructed the phylogenetic relationships of Australian Prioninae using mitochondrial protein-coding genes. Based on the molecular phylogeny and morphological characters, we suggest downgrading the subfamily Parandrinae Blanchard to the tribe Parandrini in the Prioninae Latreille and synonymise Erichsoniini Thomson with Parandrini Blanchard. Australian Prioninae are classified in seven tribes: Catypnini Lacordaire stat. rev., Mocrotomini Thomson, Osphryonini trib. nov., Parandrini Blanchard, Rhipidocerini trib. nov., Sceleocanthini Lacordaire stat. rev., and Tereticini Lameere. Tribal compositions and generic relationships are tested, and the subtribal divisions within the large tribe Macrotomini are briefly discussed.

A total of 174 pinned specumens (57 from Australia and 117 from outside Australia) were sampled for this study. DNA was extracted from leg tissue using the DNeasy 96 Blood and Tissue Kit protocol (Qiagen) and measured with a Qubit fluorometer (Thermo Fisher). Libraries were prepared following the QIAseq FX DNA Library Kit protocol (Qiagen) with one-thrid reaction volumes. The sample DNA was fragmented with the FX Enhancer for 15 minutes, and a right-side size selection was performed at the adaptor ligation cleanup stage. Samples were triple-indexed using custom adapters based on standard Illumina TruSeq UDI adaptors. Individual libraries were pooled equally and sequenced on a NovaSeq6000 S1 with 150 bp PE at the Biomolecular Resource Facility, Australian National University.

Sequencing raw reads were trimmed by Trimmomatic v.0.36 and de novo assembled using Spades v.3.11.1. Putative mitochondrial (mt) contigs were recognized using BLAST and a reference dataset containing 25 complete mt-genomes of Australian prionines, then blasted against NCBI GenBank and against each other to remove contaminants. mt-PCGs were annotated based on references and extracted accordingly, aligned at the translated amino acid level using Muscle and concatenated to a final supermatrix. Final nucleotide sequence (NT) and amino acid sequence (AA) alignments were partitioned by gene positions and analysed in ModelFinder with +MERGE option. Phylogenetic relationships were inferred under Maximum Likelihood criterion in IQTREE2 v.2.0.3. We performed 20 independent tree searches for AA and NT datasets, respectively, and selected the tree with the best likelihood score. Nodal support values were estimated using the slow standard nonparametric bootstrap in IQ-TREE2 with 200 replicates.