Noninvasive reconstruction of complete mitochondrial genomes from aquatic environmental DNA using PCR-free long-read sequencing

Environmental DNA (eDNA) is a powerful, noninvasive data source for assessing aquatic biodiversity, but the assumption that it is present only in low concentrations and highly fragmented forms has limited its use. Accordingly, eDNA genomic analyses have focused on short mitochondrial regions, which often lack the resolution for accurate species identification or individual-level analysis. However, some intact mitochondrial genomes may persist in aquatic environments, and their recovery could enable noninvasive determination of complete mitochondrial sequences, providing higher-resolution genetic information. Nevertheless, most eDNA studies have overlooked long-read DNA analysis, and current protocols are not optimised to detect or utilise these longer fragments.

To address this, we developed a long-read sequencing workflow to reconstruct complete mitochondrial genomes from eDNA. First, we selected the optimal DNA extraction method and filter pore sizes using water from aquaria containing Carassius auratus or Oryzias latipes , aiming to maximise the yield of high-molecular-weight DNA for target organism sequences. We then applied two sequencing strategies: a PCR-free approach using native eDNA and a long-range PCR-based approach targeting the complete mitochondrial genome.

Using the determined DNA extraction method and a filter (1.2 µm pore size), we obtained high-molecular-weight DNA suitable for long-read sequencing. In the PCR-free approach, C. auratus samples from single-fish aquaria yielded hundreds of reads mapped to the mitochondrial genome, with some reads exceeding 16,000 bp. These allowed de novo assembly of full-length mitochondrial genomes in several samples. In the PCR-based approach, one sample yielded over 100,000 mitochondrial reads, of which >27% exceeded 16,000 bp, enabling assemblies. Comparison with assemblies from tissue-derived DNA confirmed accuracy, as sequences were nearly identical. Across samples, polymorphism analysis revealed two distinct haplotypes in C. auratus . These results demonstrate that full-length mitochondrial genomes can be reconstructed from eDNA, and that both PCR-free and PCR-based strategies are effective under appropriate sample conditions.

Our results show that long-read sequencing enables the recovery of near-complete mitochondrial genomes directly from eDNA without requiring artificial fragmentation and, in some cases, without the need for enrichment steps. This approach enhances eDNA-based biodiversity monitoring by improving species identification and enabling within-population level analyses.