Using Next-Generation Sequencing for DNA Barcoding: Capturing Allelic Variation in ITS2

Abstract

Internal Transcribed Spacer 2 (ITS2) is a popular DNA barcoding marker; however, in some animal species it is hypervariable and therefore difficult to sequence with traditional methods. With next-generation sequencing (NGS) it is possible to sequence all gene variants despite the presence of single nucleotide polymorphisms (SNPs), insertions/deletions (indels), homopolymeric regions, and microsatellites. Our aim was to compare the performance of Sanger sequencing and NGS amplicon sequencing in characterizing ITS2 in 26 mosquito species represented by 88 samples. The suitability of ITS2 as a DNA barcoding marker for mosquitoes, and its allelic diversity in individuals and species, was also assessed. Compared to Sanger sequencing, NGS was able to characterize the ITS2 region to a greater extent, with resolution within and between individuals and species that was previously not possible. A total of 382 unique sequences (alleles) were generated from the 88 mosquito specimens, demonstrating the diversity present that has been overlooked by traditional sequencing methods. Multiple indels and microsatellites were present in the ITS2 alleles, which were often specific to species or genera, causing variation in sequence length. As a barcoding marker, ITS2 was able to separate all of the species, apart from members of the Culex pipiens complex, providing the same resolution as the commonly used Cytochrome Oxidase I (COI). The ability to cost-effectively sequence hypervariable markers makes NGS an invaluable tool with many applications in the DNA barcoding field, and provides insights into the limitations of previous studies and techniques.

Keywords: Culicidae; NGS; amplicon sequencing; indels; microsatellites.

Figure 1

ITS2 sequence information. (A) Comparison of mean Sanger sequence length post-trimming (red) and mean NGS allele length (blue) between species with standard error shown. (B) Number of alleles per individual shown as a histogram. (C) Number of alleles per species shown as a histogram. NGS, next-generation sequencing.

Figure 2

Comparison of sequences generated by Sanger and NGS technology in ITS2 regions containing different types of polymorphisms observed in this study. (A) Single nucleotide polymorphisms, (B) homopolymer region, (C) dinucleotide repeat, (D) insertion/deletion, and (E) combination. The multiple upper sequences show alleles derived from NGS data, while the lower chromatogram and single accompanying sequence were produced through Sanger sequencing. B and C are examples of microsatellites. NGS, next-generation sequencing.

Figure 3

The location of microsatellites within 26 ITS2 sequences (shown from 100 bp), each representative of a species. The size of the light blue labels are proportionate to repeat length, whereas the number inside the label is representative of the repeat type. Microsatellites within all 382 alleles can be found in Figure S2.

Figure 4

A summarized neighbor-joining tree, with bootstrap support values (%), based on p-distance comparisons between 382 ITS2 contig sequences from 88 mosquito samples. Tribal groups are indicated by brackets.

Figure 5

Distribution of percentage difference (p-distances) for ITS2 in different taxonomic categories. Comparison of conspecific and congeneric differences of (A) all 88 samples, and (B) all samples excluding those from the Dobrotworskyius genus. (C) Differences between genera in 88 samples, with peaks corresponding to the differences within and between mosquito tribes, and between mosquito subfamilies. Brackets indicate the range of p-distances for each group.