Background Mitochondrial genome sequences have become critical to the study of biodiversity. Genome skimming and other short-read based methods are the most common approaches, but they are not well-suited to scale up to multiplexing hundreds of samples. Here, we report on a new approach to sequence hundreds to thousands of complete mitochondrial genomes in parallel using long-amplicon sequencing. We amplified the mitochondrial genome of 677 specimens in two partially overlapping amplicons and implemented an asymmetric PCR-based indexing approach to multiplex 1,159 long amplicons together on a single PacBio SMRT Sequel II cell. We also tested this method on Oxford Nanopore Technologies (ONT) MinION R9.4 to assess if this method could be applied to other long-read technologies. We implemented several optimizations that make this method significantly more efficient than alternative mitochondrial genome sequencing methods. Results With the PacBio sequencing data we recovered at least one of the two fragments for 96% of samples (~ 80–90%) with mean coverage ~ 1,500x. The ONT data recovered less than 50% of input fragments likely due to low throughput and the design of the Barcoded Universal Primers which were optimized for PacBio sequencing. We compared a single mitochondrial gene alignment to half and full mitochondrial genomes and found, as expected, increased tree support with longer alignments, though whole mitochondrial genomes were not significantly better than half mitochondrial genomes. Conclusions This method can effectively capture thousands of long amplicons in a single run and be used to build more robust phylogenies quickly and effectively. We provide several recommendations for future users depending on the evolutionary scale of their system. A natural extension of this method is to collect multi-locus datasets consisting of mitochondrial genomes and several long nuclear loci at once.

Background: Mitochondrial genome sequences have become critical to the study of biodiversity, though the cost of doing so with genome skimming or other methods has been difficult to reduce beyond ~$15–20 per sample making it costly to scale up. Here, we report on an approach to sequence hundreds to thousands of complete mitochondrial genomes in parallel using long-amplicon sequencing. We amplified the mitochondrial genome of 677 specimens in two partially overlapping amplicons and implemented an asymmetric PCR-based indexing approach to multiplex 1,159 long amplicons together on a single PacBio SMRT cell. We also tested this method on Oxford Nanopore Technologies (ONT) MinION as a proof-of-concept. We implemented several cost-saving measures to reach a per-sample cost as low as $7 per mtGenome on the PacBio platform. Results: With the PacBio data we recovered at least one of the two fragments for 96% of samples (~80% of input amplicons) with mean coverage ~1,500x. The ONT data recovered less than 50% of input fragments due to the use of PacBio Barcoded Universal Primers (BUPs) and low throughput, though we are confident results will improve with ONT-optimized BUPs. We compared a single mitochondrial gene alignment to half and full mitochondrial genomes and found, as expected, significant improvement in tree support with longer alignments. Conclusions: This method can effectively capture thousands of long amplicons in a single run and be used to build more robust phylogenies quickly and effectively. We provide several recommendations for future users depending on the evolutionary scale of their system. A natural extension of this method is to collect multi-locus datasets consisting of mitochondrial genomes and several long nuclear loci at once.