Limits...
Environmental monitoring using next generation sequencing: rapid identification of macroinvertebrate bioindicator species.

Carew ME, Pettigrove VJ, Metzeling L, Hoffmann AA - Front. Zool. (2013)

Bottom Line: We find that 454 generated COI sequences successfully identified up to 96% of species in samples, but this increased up to 99% when combined with CytB sequences.We also found a strong quantitative relationship between the number of 454 sequences and individuals showing that it may be possible to estimate the abundance of species from 454 pyrosequencing data.Next generation sequencing using two genes was successful for identifying chironomid species.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Zoology, Victorian Centre for Aquatic Pollution Identification and Management (CAPIM), The University of Melbourne, Victoria 3010, Australia. mecarew@unimelb.edu.au.

ABSTRACT

Introduction: Invertebrate communities are central to many environmental monitoring programs. In freshwater ecosystems, aquatic macroinvertebrates are collected, identified and then used to infer ecosystem condition. Yet the key step of species identification is often not taken, as it requires a high level of taxonomic expertise, which is lacking in most organizations, or species cannot be identified as they are morphologically cryptic or represent little known groups. Identifying species using DNA sequences can overcome many of these issues; with the power of next generation sequencing (NGS), using DNA sequences for routine monitoring becomes feasible.

Results: In this study, we test if NGS can be used to identify species from field-collected samples in an important bioindicator group, the Chironomidae. We show that Cytochrome oxidase I (COI) and Cytochrome B (CytB) sequences provide accurate DNA barcodes for chironomid species. We then develop a NGS analysis pipeline to identifying species using megablast searches of high quality sequences generated using 454 pyrosequencing against comprehensive reference libraries of Sanger-sequenced voucher specimens. We find that 454 generated COI sequences successfully identified up to 96% of species in samples, but this increased up to 99% when combined with CytB sequences. Accurate identification depends on having at least five sequences for a species; below this level species not expected in samples were detected. Incorrect incorporation of some multiplex identifiers (MID's) used to tag samples was a likely cause, and most errors could be detected when using MID tags on forward and reverse primers. We also found a strong quantitative relationship between the number of 454 sequences and individuals showing that it may be possible to estimate the abundance of species from 454 pyrosequencing data.

Conclusions: Next generation sequencing using two genes was successful for identifying chironomid species. However, when detecting species from 454 pyrosequencing data sets it was critical to include known individuals for quality control and to establish thresholds for detecting species. The NGS approach developed here can lead to routine species-level diagnostic monitoring of aquatic ecosystems.

No MeSH data available.


Related in: MedlinePlus

Bootstrapped Kirma-2-parameter trees examining the genetic distance between the species found in this study. Neighbour joining trees are based on the 46 chironomid species that occurred at the ten field sites for two gene regions a) COI b) CytB used in this study. Both trees are construct using the same regions used to identify species in the 454 pyrosequencing experiments (395 bps for COI and 343 bps for CytB) and show the level of intraspecific variation (represented by black triangles) based on sequences from up to ten individuals pre species (the number of individuals is given in parentheses) from our DNA reference libraries.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3750358&req=5

Figure 1: Bootstrapped Kirma-2-parameter trees examining the genetic distance between the species found in this study. Neighbour joining trees are based on the 46 chironomid species that occurred at the ten field sites for two gene regions a) COI b) CytB used in this study. Both trees are construct using the same regions used to identify species in the 454 pyrosequencing experiments (395 bps for COI and 343 bps for CytB) and show the level of intraspecific variation (represented by black triangles) based on sequences from up to ten individuals pre species (the number of individuals is given in parentheses) from our DNA reference libraries.

Mentions: Identification of chironomid samples from the ten field sites indicated 46 chironomid species from three subfamilies (Table 1). Diversity of species ranged from 7 to 14 per site, identified from 32 to 167 individuals collected per site, with a total of 768 individuals collected overall. While 26 species could be identified, the remaining 20 species represented new or known species that could not be identified using only larval keys. These species are denoted as sp.‘x’. Neighbour joining trees for COI and CytB based on up to ten sequences per species for the shorter ‘454 sized’ amplicons showed all species formed distinct groups and these groups were supported by high bootstraps (Figure 1). Mean intraspecific nucleotide variation within species ranged from 0–4.2% for COI and 0–4.4% for CytB, while mean inter-specific variation ranged from 7–34.1% for CytB and 8.7-34.1% for COI, also indicating that the 454 COI and CytB amplicons were suitable for separating species. GenBank accession numbers for these sequences are given in Additional file 1: Table S1.


Environmental monitoring using next generation sequencing: rapid identification of macroinvertebrate bioindicator species.

Carew ME, Pettigrove VJ, Metzeling L, Hoffmann AA - Front. Zool. (2013)

Bootstrapped Kirma-2-parameter trees examining the genetic distance between the species found in this study. Neighbour joining trees are based on the 46 chironomid species that occurred at the ten field sites for two gene regions a) COI b) CytB used in this study. Both trees are construct using the same regions used to identify species in the 454 pyrosequencing experiments (395 bps for COI and 343 bps for CytB) and show the level of intraspecific variation (represented by black triangles) based on sequences from up to ten individuals pre species (the number of individuals is given in parentheses) from our DNA reference libraries.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3750358&req=5

Figure 1: Bootstrapped Kirma-2-parameter trees examining the genetic distance between the species found in this study. Neighbour joining trees are based on the 46 chironomid species that occurred at the ten field sites for two gene regions a) COI b) CytB used in this study. Both trees are construct using the same regions used to identify species in the 454 pyrosequencing experiments (395 bps for COI and 343 bps for CytB) and show the level of intraspecific variation (represented by black triangles) based on sequences from up to ten individuals pre species (the number of individuals is given in parentheses) from our DNA reference libraries.
Mentions: Identification of chironomid samples from the ten field sites indicated 46 chironomid species from three subfamilies (Table 1). Diversity of species ranged from 7 to 14 per site, identified from 32 to 167 individuals collected per site, with a total of 768 individuals collected overall. While 26 species could be identified, the remaining 20 species represented new or known species that could not be identified using only larval keys. These species are denoted as sp.‘x’. Neighbour joining trees for COI and CytB based on up to ten sequences per species for the shorter ‘454 sized’ amplicons showed all species formed distinct groups and these groups were supported by high bootstraps (Figure 1). Mean intraspecific nucleotide variation within species ranged from 0–4.2% for COI and 0–4.4% for CytB, while mean inter-specific variation ranged from 7–34.1% for CytB and 8.7-34.1% for COI, also indicating that the 454 COI and CytB amplicons were suitable for separating species. GenBank accession numbers for these sequences are given in Additional file 1: Table S1.

Bottom Line: We find that 454 generated COI sequences successfully identified up to 96% of species in samples, but this increased up to 99% when combined with CytB sequences.We also found a strong quantitative relationship between the number of 454 sequences and individuals showing that it may be possible to estimate the abundance of species from 454 pyrosequencing data.Next generation sequencing using two genes was successful for identifying chironomid species.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Zoology, Victorian Centre for Aquatic Pollution Identification and Management (CAPIM), The University of Melbourne, Victoria 3010, Australia. mecarew@unimelb.edu.au.

ABSTRACT

Introduction: Invertebrate communities are central to many environmental monitoring programs. In freshwater ecosystems, aquatic macroinvertebrates are collected, identified and then used to infer ecosystem condition. Yet the key step of species identification is often not taken, as it requires a high level of taxonomic expertise, which is lacking in most organizations, or species cannot be identified as they are morphologically cryptic or represent little known groups. Identifying species using DNA sequences can overcome many of these issues; with the power of next generation sequencing (NGS), using DNA sequences for routine monitoring becomes feasible.

Results: In this study, we test if NGS can be used to identify species from field-collected samples in an important bioindicator group, the Chironomidae. We show that Cytochrome oxidase I (COI) and Cytochrome B (CytB) sequences provide accurate DNA barcodes for chironomid species. We then develop a NGS analysis pipeline to identifying species using megablast searches of high quality sequences generated using 454 pyrosequencing against comprehensive reference libraries of Sanger-sequenced voucher specimens. We find that 454 generated COI sequences successfully identified up to 96% of species in samples, but this increased up to 99% when combined with CytB sequences. Accurate identification depends on having at least five sequences for a species; below this level species not expected in samples were detected. Incorrect incorporation of some multiplex identifiers (MID's) used to tag samples was a likely cause, and most errors could be detected when using MID tags on forward and reverse primers. We also found a strong quantitative relationship between the number of 454 sequences and individuals showing that it may be possible to estimate the abundance of species from 454 pyrosequencing data.

Conclusions: Next generation sequencing using two genes was successful for identifying chironomid species. However, when detecting species from 454 pyrosequencing data sets it was critical to include known individuals for quality control and to establish thresholds for detecting species. The NGS approach developed here can lead to routine species-level diagnostic monitoring of aquatic ecosystems.

No MeSH data available.


Related in: MedlinePlus