Inferring species membership using DNA sequences with back-propagation neural networks

DNA barcoding as a method for species identification is rapidly increasing in popularity. However, there are still relatively few rigorous methodological tests of DNA barcoding. Current distance-based methods are frequently criticized for treating the nearest neighbor as the closest relative via a raw similarity score, lacking an objective set of criteria to delineate taxa, or for being incongruent with classical character-based taxonomy. Here, we propose an artificial intelligence-based approach - inferring species membership via DNA barcoding with back-propagation neural networks (named BP-based species identification) - as a new advance to the spectrum of available methods. We demonstrate the value of this approach with simulated data sets representing different levels of sequence variation under coalescent simulations with various evolutionary models, as well as with two empirical data sets of COI sequences from East Asian ground beetles (Carabidae) and Costa Rican skipper butterflies. With a 630-to 690-bp fragment of the COI gene, we identified 97.50% of 80 unknown sequences of ground beetles, 95.63%, 96.10%, and 100% of 275, 205, and 9 unknown sequences of the neotropical skipper butterfly to their correct species, respectively. Our simulation studies indicate that the success rates of species identification depend on the divergence of sequences, the length of sequences, and the number of reference sequences. Particularly in cases involving incomplete lineage sorting, this new BP-based method appears to be superior to commonly used methods for DNA-based species identification.     

DNA barcoding of Murinae (Rodentia: Muridae) and Arvicolinae (Rodentia: Cricetidae) distributed in China

Identification of rodents is very difficult mainly due to high similarities in morphology and controversial taxonomy. In this study, mitochondrial cytochrome oxidase subunit I (COI) was used as DNA barcode to identify the Murinae and Arvicolinae species distributed in China and to facilitate the systematics studies of Rodentia. In total, 242 sequences (31 species, 11 genera) from Murinae and 130 sequences (23 species, 6 genera) from Arvicolinae were investigated, of which 90 individuals were novel. Genetic distance, threshold method, tree‐based method, online BLAST and BLOG were employed to analyse the data sets. There was no obvious barcode gap. The average K2P distance within species and genera was 2.10% and 12.61% in Murinae, and 2.86% and 11.80% in Arvicolinae, respectively. The optimal threshold was 5.62% for Murinae and 3.34% for Arvicolinae. All phylogenetic trees exhibited similar topology and could distinguish 90.32% of surveyed species in Murinae and 82.60% in Arvicolinae with high support values. BLAST analyses yielded similar results with identification success rates of 92.15% and 93.85% for Murinae and Arvicolinae, respectively. BLOG successfully authenticated 100% of detected species except Leopoldamys edwardsi based on the latest taxonomic revision. Our results support the species status of recently recognized Micromys erythrotis, Eothenomys tarquinius and E. hintoni and confirm the important roles of comprehensive taxonomy and accurate morphological identification in DNA barcoding studies. We believe that, when proper analytic methods are applied or combined, DNA barcoding could serve as an accurate and effective species identification approach for Murinae and Arvicolinae based on a proper taxonomic framework.     

Phylogenetic reconstruction and DNA barcoding for closely related pine moth species (Dendrolimus) in China with multiple gene markers

Unlike distinct species, closely related species offer a great challenge for phylogeny reconstruction and species identification with DNA barcoding due to their often overlapping genetic variation. We tested a sibling species group of pine moth pests in China with a standard cytochrome c oxidase subunit I (COI) gene and two alternative internal transcribed spacer (ITS) genes (ITS1 and ITS2). Five different phylogenetic/DNA barcoding analysis methods (Maximum likelihood (ML)/Neighbor-joining (NJ), "best close match" (BCM), Minimum distance (MD), and BP-based method (BP)), representing commonly used methodology (tree-based and non-tree based) in the field, were applied to both single-gene and multiple-gene analyses. Our results demonstrated clear reciprocal species monophyly for three relatively distant related species, Dendrolimus superans, D. houi, D. kikuchii, as recovered by both single and multiple genes while the phylogenetic relationship of three closely related species, D. punctatus, D. tabulaeformis, D. spectabilis, could not be resolved with the traditional tree-building methods. Additionally, we find the standard COI barcode outperforms two nuclear ITS genes, whatever the methods used. On average, the COI barcode achieved a success rate of 94.10-97.40%, while ITS1 and ITS2 obtained a success rate of 64.70-81.60%, indicating ITS genes are less suitable for species identification in this case. We propose the use of an overall success rate of species identification that takes both sequencing success and assignation success into account, since species identification success rates with multiple-gene barcoding system were generally overestimated, especially by tree-based methods, where only successfully sequenced DNA sequences were used to construct a phylogenetic tree. Non-tree based methods, such as MD, BCM, and BP approaches, presented advantages over tree-based methods by reporting the overall success rates with statistical significance. In addition, our results indicate that the most closely related species D. punctatus, D. tabulaeformis, and D. spectabilis, may be still in the process of incomplete lineage sorting, with occasional hybridizations occurring among them.     

A comparison of algorithms for the identification of specimens using DNA barcodes: examples from gymnosperms

In order to use DNA sequences for specimen identification (e.g., barcoding, fingerprinting) an algorithm to compare query sequences with a reference database is needed. Precision and accuracy of query sequence identification was estimated for hierarchical clustering (parsimony and neighbor joining), similarity methods (BLAST, BLAT and megaBLAST), combined clustering/similarity methods (BLAST/parsimony and BLAST/neighbor joining), diagnostic methods (DNA–BAR and DOME ID), and a new method (ATIM). We offer two novel alignment‐free algorithmic solutions (DOME ID and ATIM) to identify query sequences for the purposes of DNA barcoding. Publicly available gymnosperm nrITS 2 and plastid matK sequences were used as test data sets. On the test data sets, almost all of the methods were able to accurately identify sequences to genus; however, no method was able to accurately identify query sequences to species at a frequency that would be considered useful for routine specimen identification (42–71% unambiguously correct). Clustering methods performed the worst (perhaps due to alignment issues). Similarity methods, ATIM, DNA–BAR, and DOME ID all performed at approximately the same level. Given the relative precision of the algorithms (median = 67% unambiguous), the low accuracy of species‐level identification observed could be ascribed to the lack of correspondence between patterns of allelic similarity and species delimitations. Application of DNA barcoding to sequences of CITES listed cycads (Cycadopsida) provides an example of the potential application of DNA barcoding to enforcement of conservation laws. © The Willi Hennig Society 2006.     

以雾灵山夜蛾科为材料评估进化模型对DNA条码分类的影响

为了探究进化模型对DNA条形码分类的影响, 本研究以雾灵山夜蛾科44个种的标本为材料, 获得COI基因序列。使用邻接法（neighbor-joining）、 最大简约法（maximum parsimony）、 最大似然法（maximum likelihood）以及贝叶斯法（Bayesian inference）构建系统发育树, 并且对邻接法的12种模型、 最大似然法的7种模型、 贝叶斯法的2种模型进行模型成功率的评估。结果表明, 邻接法的12种模型成功率相差不大, 较稳定; 最大似然法及贝叶斯法的不同模型成功率存在明显差异, 不稳定; 最大简约法不基于模型, 成功率比较稳定。邻接法及最大似然法共有6种相同的模型, 这6种模型在不同的方法中成功率存在差异。此外, 分子数据中存在单个物种仅有一条序列的情况, 显著降低了模型成功率, 表明在DNA条形码研究中, 每个物种需要有多个样本。     

Testing the reliability of genetic methods of species identification via simulation

Although genetic methods of species identification, especially DNA barcoding, are strongly debated, tests of these methods have been restricted to a few empirical cases for pragmatic reasons. Here we use simulation to test the performance of methods based on sequence comparison (BLAST and genetic distance) and tree topology over a wide range of evolutionary scenarios. Sequences were simulated on a range of gene trees spanning almost three orders of magnitude in tree depth and in coalescent depth; that is, deep or shallow trees with deep or shallow coalescences. When the query's conspecific sequences were included in the reference alignment, the rate of positive identification was related to the degree to which different species were genetically differentiated. The BLAST, distance, and liberal tree-based methods returned higher rates of correct identification than did the strict tree-based requirement that the query was within, but not sister to, a single-species clade. Under this more conservative approach, ambiguous outcomes occurred in inverse proportion to the number of reference sequences per species. When the query's conspecific sequences were not in the reference alignment, only the strict tree-based approach was relatively immune to making false-positive identifications. Thresholds affected the rates at which false-positive identifications were made when the query's species was unrepresented in the reference alignment but did not otherwise influence outcomes. A conservative approach using the strict tree-based method should be used initially in large-scale identification systems, with effort made to maximize sequence sampling within species. Once the genetic variation within a taxonomic group is well characterized and the taxonomy resolved, then the choice of method used should be dictated by considerations of computational efficiency. The requirement for extensive genetic sampling may render these techniques inappropriate in some circumstances.     

Testing four proposed barcoding markers for the identification of species within Ligustrum L.(Oleaceae)

DNA barcoding is a biological technique that uses short and standardized genes or DNA regions to facilitate species identification. DNA barcoding has been used successfully in several animal and plant groups. Ligustrum (Oleaceae) species occur widely throughout the world and are used as medicinal plants in China. Therefore, the accurate identification of species in this genus is necessary. Four potential DNA barcodes, namely the nuclear ribosomal internal transcribed spacer (ITS) and three chloroplast (cp) DNA regions (rbcL, marK, and trnH-psbA),were used to differentiate species within Ligustrum. BLAST, character-based method, tree-based methods and TAXONDNA analysis were used to investigate the molecular identification capabilities of the chosen markers for discriminating 92 samples representing 20 species of this genus. The results showed that the ITS sequences have the most variable information, followed by trnH-psbA, matK, and rbcL. All sequences of the four regions correctly identified the species at the genus level using BLAST alignment. At the species level, the discriminating power of rbcL, matK, trnH-psbA and ITS based on neighbor-joining (NJ) trees was 36.8%, 38.9%, 77.8%, and 80%,respectively. Using character-based and maximum parsimony (MP) tree methods together, the discriminating ability of trnH-psbA increased to 88.9%. All species could be differentiated using ITS when combining the NJ tree method with character-based or MP tree methods. Overall, the results indicate that DNA barcoding is an effective molecular identification method for Ligustrum species. We propose the nuclear ribosomal ITS as a plant barcode for plant identification and trnH-psbA as a candidate barcode sequence.     

DNA barcoding Central Asian butterflies: increasing geographical dimension does not significantly reduce the success of species identification

DNA barcoding employs short, standardized gene regions (5' segment of mitochondrial cytochrome oxidase subunit I for animals) as an internal tag to enable species identification. Prior studies have indicated that it performs this task well, because interspecific variation at cytochrome oxidase subunit I is typically much greater than intraspecific variation. However, most previous studies have focused on local faunas only, and critics have suggested two reasons why barcoding should be less effective in species identification when the geographical coverage is expanded. They suggested that many recently diverged taxa will be excluded from local analyses because they are allopatric. Second, intraspecific variation may be seriously underestimated by local studies, because geographical variation in the barcode region is not considered. In this paper, we analyse how adding a geographical dimension affects barcode resolution, examining 353 butterfly species from Central Asia. Despite predictions, we found that geographically separated and recently diverged allopatric species did not show, on average, less sequence differentiation than recently diverged sympatric taxa. Although expanded geographical coverage did substantially increase intraspecific variation reducing the barcoding gap between species, this did not decrease species identification using neighbour-joining clustering. The inclusion of additional populations increased the number of paraphyletic entities, but did not impede species-level identification, because paraphyletic species were separated from their monophyletic relatives by substantial sequence divergence. Thus, this study demonstrates that DNA barcoding remains an effective identification tool even when taxa are sampled from a large geographical area.     

An evaluation of candidate plant DNA barcodes and assignment methods in diagnosing 29 species in the genus Agalinis (Orobanchaceae)

? Premise of the study: DNA barcoding has been proposed as a useful technique within many disciplines (e.g., conservation biology and forensics) for determining the taxonomic identity of a sample based on nucleotide similarity to samples of known taxonomy. Application of DNA barcoding to plants has primarily focused on evaluating the success of candidate barcodes across a broad spectrum of evolutionary divergence. Less attention has been paid to evaluating performance when distinguishing congeners or to differential success of analytical techniques despite the fact that the practical application and utility of barcoding hinges on the ability to distinguish closely related species. ? Methods: We tested the ability to distinguish among 92 samples representing 29 putative species in the genus Agalinis (Orobanchaceae) using 13 candidate barcodes and three analytical methods (i.e., threshold genetic distances, hierarchical tree-based, and diagnostic character differences). Due to questions regarding evolutionary distinctiveness of some taxa, we evaluated success under two taxonomic hypotheses. ? Key results: The psbA-trnH and trnT-trnL barcodes in conjunction with the "best close match" distance-based method best met the objectives of DNA barcoding. Success was also a function of the taxonomy used. ? Conclusions: In addition to accurately identifying query sequences, our results showed that DNA barcoding is useful for detecting taxonomic uncertainty; determining whether erroneous taxonomy or incomplete lineage sorting is the cause requires additional information provided by traditional taxonomic approaches. The magnitude of differentiation within and among the Agalinis species sampled suggests that our results inform how DNA barcoding will perform among closely related species in other genera.     

Testing four proposed barcoding markers for the identification of species within Ligustrum L. (Oleaceae)

DNA barcoding is a biological technique that uses short and standardized genes or DNA regions to facilitate species identification. DNA barcoding has been used successfully in several animal and plant groups. Ligustrum (Oleaceae) species occur widely throughout the world and are used as medicinal plants in China. Therefore, the accurate identification of species in this genus is necessary. Four potential DNA barcodes, namely the nuclear ribosomal internal transcribed spacer (ITS) and three chloroplast (cp) DNA regions (rbcL, matK, and trnH–psbA), were used to differentiate species within Ligustrum. BLAST, character-based method, tree-based methods and TAXONDNA analysis were used to investigate the molecular identification capabilities of the chosen markers for discriminating 92 samples representing 20 species of this genus. The results showed that the ITS sequences have the most variable information, followed by trnH–psbA, matK, and rbcL. All sequences of the four regions correctly identified the species at the genus level using BLAST alignment. At the species level, the discriminating power of rbcL, matK, trnH–psbA, and ITS based on neighbor-joining (NJ) trees was 36.8%, 38.9%, 77.8%, and 80%, respectively. Using character-based and maximum parsimony (MP) tree methods together, the discriminating ability of trnH–psbA increased to 88.9%. All species could be differentiated using ITS when combining the NJ tree method with character-based or MP tree methods. Overall, the results indicate that DNA barcoding is an effective molecular identification method for Ligustrum species. We propose the nuclear ribosomal ITS as a plant barcode for plant identification and trnH–psbA as a candidate barcode sequence.     

DNA barcoding of Cymbidium by genome skimming: Call for next-generation nuclear barcodes

Cymbidium is an orchid genus that has undergone rapid radiation and has high ornamental, economic, ecological and cultural importance, but its classification based on morphology is controversial. The plastid genome (plastome), as an extension of plant standard DNA barcodes, has been widely used as a potential molecular marker for identifying recently diverged species or complicated plant groups. In this study, we newly generated 237 plastomes of 50 species (at least two individuals per species) by genome skimming, covering 71.4% of members of the genus Cymbidium. Sequence-based analyses (barcoding gaps and automatic barcode gap discovery) and tree-based analyses (maximum likelihood, Bayesian inference and multirate Poisson tree processes model) were conducted for species identification of Cymbidium. Our work provides a comprehensive DNA barcode reference library for Cymbidium species identification. The results show that compared with standard DNA barcodes (rbcL + matK) as well as the plastid trnH-psbA, the species identification rate of the plastome increased moderately from 58% to 68%. At the same time, we propose an optimized identification strategy for Cymbidium species. The plastome cannot completely resolve the species identification of Cymbidium, the main reasons being incomplete lineage sorting, artificial cultivation, natural hybridization and chloroplast capture. To further explore the potential use of nuclear data in identifying species, the Skmer method was adopted and the identification rate increased to 72%. It appears that nuclear genome data have a vital role in species identification and are expected to be used as next-generation nuclear barcodes.     

DNA barcoding in medicinal plants: Testing the potential of a proposed barcoding marker for identification of Uncaria species from China

DNA barcoding, an increasingly popular mean of species identification, has been widely used for global species identification despite a consensus not being reached regarding which DNA sequences can be used as the best plant barcodes. In this study, we tested the feasibility of five candidate DNA barcodes (nrITS, nrITS2, matk, rbcL and trnH-psbA) for identifying Uncaria species. We collected a total of 54 specimens of 10 Uncaria species across its distributional range. BLAST, barcoding gaps, tree-based methods and TAXONDNA analysis were used to investigate the molecular identification capability of the candidate DNA barcodes. The results showed that the ITS2 is most suitable as a candidate DNA barcode for identification of medicinal plants of the genus Uncaria.     

DNA barcodes and microsatellites: How they complement for species identification in the complex genus Tamarix (Tamaricaceae)

DNA barcoding allows the identification of an organism by comparing the sequence of selected DNA regions (barcodes) with a previously compiled database, and it can be useful for taxonomic identification of species in complex genera, such as Tamarix. Many species of this genus show convergent morphology, which leads to frequent errors in their identification. Highly variable genetic markers, such as microsatellites or short sequence repeats (SSR), could be used to differentiate species where DNA barcodes fail. Here, we tested the ability of both, 5 different marker regions (rbcL, matK, ITS, trnH-psbA, and ycf1), and 14 microsatellites, to properly identify Tamarix species, especially those from the Mediterranean Basin, and compared the pros and cons of the different analytical methods for species identification. DNA barcoding allows the genetic identification of certain species in Tamarix. The two-locus barcodes matK + ITS and ITS + ycf1 were the best-performing combinations, allowing up to 69% and 70%, respectively, correct identification. However, DNA barcoding failed in phylogenetically close groups, such as many Mediterranean species. The use of SSR can aid the identification of species, and the combination of both types of data (DNA barcoding and SSR) improved the success. The combination of data was especially relevant in detecting the presence of hybridization processes, which are common in the genus. However, caution must be exercised when choosing the clustering methods for the SSR datasince different methods can lead to very different results.     

DNA barcoding meets molecular scatology: short mtDNA sequences for standardized species assignment of carnivore noninvasive samples

Although species assignment of scats is important to study carnivore biology, there is still no standardized assay for the identification of carnivores worldwide, which would allow large-scale routine assessments and reliable cross-comparison of results. Here, we evaluate the potential of two short mtDNA fragments [ATP6 (126 bp) and cytochrome oxidase I gene (COI) (187 bp)] to serve as standard markers for the Carnivora. Samples of 66 species were sequenced for one or both of these segments. Alignments were complemented with archival sequences and analysed with three approaches (tree-based, distance-based and character-based). Intraspecific genetic distances were generally lower than between-species distances, resulting in diagnosable clusters for 86% (ATP6) and 85% (COI) of the species. Notable exceptions were recently diverged species, most of which could still be identified using diagnostic characters and uniqueness of haplotypes or by reducing the geographic scope of the comparison. In silico analyses were also performed for a 110-bp cytochrome b (cytb) segment, whose identification success was lower (70%), possibly due to the smaller number of informative sites and/or the influence of misidentified sequences obtained from GenBank. Finally, we performed case studies with faecal samples, which supported the suitability of our two focal markers for poor-quality DNA and allowed an assessment of prey DNA co-amplification. No evidence of prey DNA contamination was found for ATP6, while some cases were observed for COI and subsequently eliminated by the design of more specific primers. Overall, our results indicate that these segments hold good potential as standard markers for accurate species-level identification in the Carnivora.     

Existence of species complex largely reduced barcoding success for invasive species of Tephritidae: a case study in Bactrocera spp.

Fruit flies in the family Tephritidae are the economically important pests that have many species complexes. DNA barcoding has gradually been verified as an effective tool for identifying species in a wide range of taxonomic groups, and there are several publications on rapid and accurate identification of fruit flies based on this technique; however, comprehensive analyses of large and new taxa for the effectiveness of DNA barcoding for fruit flies identification have been rare. In this study, we evaluated the COI barcode sequences for the diagnosis of fruit flies using 1426 sequences for 73 species of Bactrocera distributed worldwide. Tree‐based [neighbour‐joining (NJ)]; distance‐based, such as Best Match (BM), Best Close Match (BCM) and Minimum Distance (MD); and character‐based methods were used to evaluate the barcoding success rates obtained with maintaining the species complex in the data set, treating a species complex as a single taxon unit, and removing the species complex. Our results indicate that the average divergence between species was 14.04% (0.00–25.16%), whereas within a species this was 0.81% (0.00–9.71%); the existence of species complexes largely reduced the barcoding success for Tephritidae, for example relatively low success rates (74.4% based on BM and BCM and 84.8% based on MD) were obtained when the sequences from species complexes were included in the analysis, whereas significantly higher success rates were achieved if the species complexes were treated as a single taxon or removed from the data set – BM (98.9%), BCM (98.5%) and MD (97.5%), or BM (98.1%), BCM (97.4%) and MD (98.2%).     

The Use of DNA Barcoding on Recently Diverged Species in the Genus Gentiana (Gentianaceae) in China

DNA barcoding of plants poses particular challenges, especially in differentiating, recently diverged taxa. The genus Gentiana (Gentianaceae) is a species-rich plant group which rapidly radiated in the Himalaya-Hengduan Mountains in China. In this study, we tested the core plant barcode (rbcL + matK) and three promising complementary barcodes (trnH-psbA, ITS and ITS2) in 30 Gentiana species across 6 sections using three methods (the genetic distance-based method, Best Close Match and tree-based method). rbcL had the highest PCR efficiency and sequencing success (100%), while the lowest sequence recoverability was from ITS (68.35%). The presence of indels and inversions in trnH-psbA in Gentiana led to difficulties in sequence alignment. When using a single region for analysis, ITS exhibited the highest discriminatory power (60%-74.42%). Of the combinations, matK + ITS provided the highest discrimination success (71.43%-88.24%) and is recommended as the DNA barcode for the genus Gentiana. DNA barcoding proved effective in assigning most species to sections, though it performed poorly in some closely related species in sect. Cruciata because of hybridization events. Our analysis suggests that the status of G. pseudosquarrosa needs to be studied further. The utility of DNA barcoding was also verified in authenticating ‘Qin-Jiao’ Gentiana medicinal plants (G. macrophylla, G. crassicaulis, G. straminea, and G. dahurica), which can help ensure safe and correct usage of these well-known Chinese traditional medicinal herbs.     

Failure of DNA barcoding in discriminating Calligonum species

We tested the effectiveness of four DNA barcoding markers (rbcL, matK, ITS and trnLF region) for land plants in identifying Calligonum species. High quality sequences were obtained for rbcL, matK and trnLF with the universal primers whereas ITS sequences were of poor quality. RbcL and matK were highly conservative and failed in species discrimination. When rbcL, matK and trnLF were combined, the species resolution was up to 6.25%. Low sequence variation resulted in poorly resolved tree topologies. Among the sixteen sampled species, only three were recovered as a monophyletic group. Our results show that although DNA barcoding is an important tool for species identification, it fails in discriminating Calligonum species. Further research will be needed to develop markers capable to discriminate species in this taxonomy complicated and recently diverged genus.     

Comparing the usefulness of distance, monophyly and character-based DNA barcoding methods in species identification: a case study of neogastropoda

Background

DNA barcoding has recently been proposed as a promising tool for the rapid species identification in a wide range of animal taxa. Two broad methods (distance and monophyly-based methods) have been used. One method is based on degree of DNA sequence variation within and between species while another method requires the recovery of species as discrete clades (monophyly) on a phylogenetic tree. Nevertheless, some issues complicate the use of both methods. A recently applied new technique, the character-based DNA barcode method, however, characterizes species through a unique combination of diagnostic characters.

Methodology/Principal Findings

Here we analyzed 108 COI and 102 16S rDNA sequences of 40 species of Neogastropoda from a wide phylogenetic range to assess the performance of distance, monophyly and character-based methods of DNA barcoding. The distance-based method for both COI and 16S rDNA genes performed poorly in terms of species identification. Obvious overlap between intraspecific and interspecific divergences for both genes was found. The “10× rule” threshold resulted in lumping about half of distinct species for both genes. The neighbour-joining phylogenetic tree of COI could distinguish all species studied. However, the 16S rDNA tree could not distinguish some closely related species. In contrast, the character-based barcode method for both genes successfully identified 100% of the neogastropod species included, and performed well in discriminating neogastropod genera.

Conclusions/Significance

This present study demonstrates the effectiveness of the character-based barcoding method for species identification in different taxonomic levels, especially for discriminating the closely related species. While distance and monophyly-based methods commonly use COI as the ideal gene for barcoding, the character-based approach can perform well for species identification using relatively conserved gene markers (e.g., 16S rDNA in this study). Nevertheless, distance and monophyly-based methods, especially the monophyly-based method, can still be used to flag species.     

FuzzyID2: A software package for large data set species identification via barcoding and metabarcoding using hidden Markov models and fuzzy set methods

Species identification through DNA barcoding or metabarcoding has become a key approach for biodiversity evaluation and ecological studies. However, the rapid accumulation of barcoding data has created some difficulties: for instance, global enquiries to a large reference library can take a very long time. We here devise a two‐step searching strategy to speed identification procedures of such queries. This firstly uses a Hidden Markov Model (HMM) algorithm to narrow the searching scope to genus level and then determines the corresponding species using minimum genetic distance. Moreover, using a fuzzy membership function, our approach also estimates the credibility of assignment results for each query. To perform this task, we developed a new software pipeline, FuzzyID2, using Python and C++. Performance of the new method was assessed using eight empirical data sets ranging from 70 to 234,535 barcodes. Five data sets (four animal, one plant) deployed the conventional barcode approach, one used metabarcodes, and two were eDNA‐based. The results showed mean accuracies of generic and species identification of 98.60% (with a minimum of 95.00% and a maximum of 100.00%) and 94.17% (with a range of 84.40%–100.00%), respectively. Tests with simulated NGS sequences based on realistic eDNA and metabarcode data demonstrated that FuzzyID2 achieved a significantly higher identification success rate than the commonly used Blast method, and the TIPP method tends to find many fewer species than either FuzztID2 or Blast. Furthermore, data sets with tens of thousands of barcodes need only a few seconds for each query assignment using FuzzyID2. Our approach provides an efficient and accurate species identification protocol for biodiversity‐related projects with large DNA sequence data sets.