Intraspecific genetic variation in complex assemblages from mitochondrial metagenomics: comparison with DNA barcodes

1. Metagenomic shotgun sequencing, using Illumina technology, and de novo genome assembly of mixed field-collected samples of invertebrates readily produce mitochondrial genome sequences, allowing rapid identification and quantification of species diversity. However, intraspecific genetic variability present in the specimen pools is lost during mitogenome assembly, which limits the utility of mitochondrial metagenomics' for studies of population diversity. Using 10 natural communities (>2600 individuals) of leaf beetles (Chrysomelidae), DNA variation in the mitochondrial cox1-5 barcode' was compared for Sanger-sequenced individuals and Illumina shotgun-sequenced specimen pools. Generally, only a single mitochondrial contig was assembled per species, even in the presence of intraspecific variation. Ignoring ambiguity from the use of two different assemblers, the cox1 barcode regions from these assemblies were exact nucleotide matches of a Sanger-sequenced barcode in 907% of cases, which dropped to 76% in assemblies from samples with large intra- and interspecific variability. Nucleotide differences between barcodes from both data types were almost exclusively in synonymous 3rd codon positions, although the number of affected sites was very low, and the greatest discrepancies were correlated with poor quality of Sanger sequences. Unassembled shotgun reads were also used to score single nucleotide polymorphisms and to calculate intraspecific nucleotide diversity (pi) for all available populations at each site. These values correlated with Sanger-sequenced cox1 variation but were significantly higher. Overall, the assemblage-focused shotgun sequencing of pooled samples produced nucleotide variation data comparable to the well-established specimen-focused Sanger approach. The findings thus extend the application of mitochondrial metagenomics of complex biodiversity samples to the estimation of diversity below the species level.