Open Access
Translator Disclaimer
28 October 2020 Use of Morphometric Analyses and DNA Barcoding to Distinguish Truncilla donaciformis and Truncilla truncata (Bivalvia: Unionidae)
Tyler W. Beyett, Kelly McNichols-O'Rourke, Todd J. Morris, David T. Zanatta
Author Affiliations +
Abstract

Closely related unionid species often overlap in shell shape and can be difficult to accurately identify in the field. Ambiguity in identification can have serious impacts on conservation efforts and population surveys of threatened and endangered species. Truncilla donaciformis and Truncilla truncata are sister species that overlap in their distributions and frequently co-occur in central North America. Because T. donaciformis is endangered in Canada and imperiled in some US jurisdictions, co-occurrence with the morphologically similar T. truncata means that misidentification could seriously impact status assessments and recovery efforts. The objectives of this study were to (1) establish species identifications of specimens using DNA barcoding (COI), (2) determine how well traditional morphometrics and geometric morphometrics accurately discriminate between the two species, and (3) determine the accuracy of field identifications relative to molecular and morphometric identifications. Truncilla specimens from four rivers in southern Ontario were photographed and visceral mass swabs were taken. Positive identifications of all specimens were obtained through DNA barcoding and comparison with sequences from GenBank. Traditional and geometric morphometric approaches were used to assign specimens to species. Assignments generated were compared to identifications based on mtDNA barcodes, with traditional and geometric morphometric analyses found to be 90% and 99% accurate in species identifications, respectively. This study confirmed the presence of T. donaciformis in Ontario's Thames River, and revealed that all Truncilla collected for this study from the other three rivers were T. truncata. This study reinforces the utility of combining geometric morphometric analyses and DNA barcoding for identifying problematic unionid specimens.

INTRODUCTION

Early delineations and descriptions of freshwater mussel species (order Unionida) were based on shell morphology (Watters et al. 2009), which can be expressed through coloration, shape, shell sculpture, or size. Some shell-shape characters have a clear genetic basis and are potentially adaptive (Inoue et al. 2013, 2014), but habitat and environment can also have major effects on shell morphology. Use of shell shapes led to overdescription of some species based on sometimes-subtle differences (Haag 2012). More recently, species descriptions have been based on internal soft-tissue anatomy and molecular tools, resulting in the synonymizing of many previously described species (Watters et al. 2009; Haag 2012). Sacrificing animals to examine soft tissues is often not an option when dealing with threatened and endangered species, so identifications of live animals are limited to external shell features and genetic methods (e.g., DNA barcoding; Hebert et al. 2003). Because failure to correctly identify species can have important implications for the conservation of unionid diversity, a major challenge in correctly identifying species is dealing with intraspecific variation in shell shape. As shell morphology remains the most common tool used in field survey identifications of freshwater mussels, this intraspecific variation can be problematic when attempting to differentiate among closely related species with similar shell morphologies.

Truncilla donaciformis (Lea, 1828; Fawnsfoot) and Truncilla truncata (Rafinesque, 1820; Deertoe) are sister species (Burlakova et al. 2019) with similar shell morphologies (Fig. 1). Both are described mostly using shell morphology that is variable and potentially nondiagnostic. Watters et al. (2009) qualitatively describes T. donaciformis as more elongate in shape, whereas T. truncata is more triangular and typically exhibits a prominent posterior ridge. Further confounding correct field identifications, both species exhibit some subtle sexual dimorphism, with females being more rounded along the ventral margin (Watters et al. 2009; Burlakova et al. 2019).

Both T. donaciformis and T. truncata are distributed throughout much of the Mississippi River and Ohio River drainages and parts of the Great Lakes drainage of North America (Watters et al. 2009). In Canada, T. truncata is relatively common in southwestern Ontario, while T. donaciformis is considered an endangered species (COSEWIC 2008). In the USA, T. donaciformis is considered imperiled in several U.S. states and may be declining across its range (NatureServe 2020). The distribution of T. truncata is similar to T. donaciformis, but it is more widely distributed throughout the USA and is considered less imperiled. As the distributions of T. donaciformis and T. truncata are mostly sympatric (Watters et al. 2009), and their morphological characters are known to overlap, differentiating between the species can be problematic. This potential ambiguity is even more concerning when one considers the disparity in conservation status of the two species. Misidentifications could lead to misspent resources and incorrect conservation and recovery strategies, resulting in a high cost in terms of both conservation capital and species outcomes (Shea et al. 2011).

A number of practices can be employed to improve the accuracy of identifications based on shell appearance. One such practice is morphometric analysis, which measures and quantifies shape. Traditional morphometrics (TM) in mollusks include the measurement and comparison of ratios between characters such as shell length, width, and height (e.g., Cyr et al. 2007). For some species, once a large sample size of confirmed specimens has been examined, TM can be reasonably accurate in differentiating species, but positive identification cannot be assured due to morphometric variation within populations (Cyr et al. 2007; Inoue et al. 2013, 2014). If the differences are great, traditional morphometrics are usually sufficient for differentiating between groups, but if the differences are subtle, they can be confounded easily (Webster and Sheets 2010). Despite these drawbacks, taking shell measurements is quick, convenient, low-cost, and noninvasive, and these measurements are recorded in almost every field survey of unionid mussels. Geometric morphometrics (GM) involves the use of landmarks and pseudo-landmarks to assess the shape of an organism in two or three dimensions (Zelditch et al. 2012). Multivariate statistical analyses of the data are used to compare morphologies of the target specimens (Webster and Sheets 2010). Typically, these analyses are more robust and accurate than TM (Rohlf and Marcus 1993; Adams et al. 2004); their accuracy makes them an ideal method of differentiating between two occasionally ambiguous species. A reliable quantitative approach to species identification using a combination of morphometric analyses and DNA barcoding (see below) is preferable over “best guess” or “expert opinion.”

Figure 1.

Positions of geometric morphometric type I and type II landmarks on the left valve of A) typical Truncilla truncata specimen and B) typical T. donaciformis specimen. Type I (anchor) landmarks: LM 1 were placed at the tip of the umbo, LM 2 at the posterior end of the hinge ligament. Type II landmarks: LM 3 to LM 18 were placed where a fan with 40 rays (anchored at midpoint between LM 1 and LM 2) crossed the shell margin; LM 19 and 20 were placed where rays from the fan cross the anterior side of the umbo.

img-z2-5_99.jpg

In order to assess the utility and reliability of either morphometric method, it is important to first independently confirm specimen identifications. In the last two decades, DNA barcoding has become an important tool for species identification (Hebert et al. 2003; Baird and Sweeney 2011). A (∼650 bp) fragment of mitochondrial DNA encoding the cytochrome oxidase subunit 1 (COI) gene (Folmer et al. 1994) is often used as a reference sequence to which homologous sequences are compared, and barcoding with this sequence has been employed in the study of many animal groups including unionids (e.g., Inoue et al. 2014, 2018). Previously published Truncilla COI sequences (including those from Burlakova et al. 2019) on NCBI GenBank enable the positive identification of T. donaciformis and T. truncata specimens, which can help establish a group of specimens with confirmed identifications that can be used to discover morphological differences between the two species.

The objectives of this study were to (1) establish species identifications of specimens using DNA barcoding (COI), (2) determine how well TM and GM accurately discriminate between the two species, and (3) determine the accuracy of field identifications relative to molecular and morphometric identifications. We predict that GM will prove more accurate than TM in correctly differentiating between the species and that the combination of these techniques will elucidate areas of difference between the shell morphologies of T. donaciformis and T. truncata.

METHODS

We collected specimens of Truncilla donaciformis and T. truncata from the Ausable, Sydenham, Thames, and Welland rivers in Ontario, Canada, during targeted surveys completed in summer 2017 (Table 1 and Fig. 2). Swabs of the foot and mantle were taken from all live specimens and stored in lysis buffer (Sambrook et al. 1989). We photographed all field-collected specimens and returned them to the stream of origin after processing. Additional specimens were added to the morphometric dataset from the University of Michigan Museum of Zoology and the Ohio State University Division of Mollusks. The museum specimens used were lots from the Great Lakes drainage (Ontario, Michigan, and Ohio) and the Ohio River (the type locality for both species; Table 2 and Fig. 2). The left valve of each field- or museum-collected specimen was photographed in the same orientation for geometric morphometric analyses. For photographs, field-collected specimens (all live) were placed on a bed of fine sand, and a digital camera was mounted on a portable copy stand and placed at a 90° angle to the shell valve. Museum specimens (dead shell valves) were photographed using a digital camera mounted in a Stackshot apparatus with the valve placed on a piece of black fabric or modeling clay to ensure it was precisely at a 90° angle to the camera.

A 250 µL aliquot of the swab lysis buffer from each specimen was digested with 15 µL of proteinase K overnight at 56°C, followed by alcohol extraction and purification (Sambrook et al. 1989). Extracted DNA (stained with SYBR Green) was electrophoresed on a 1.5% agarose gel to confirm the success of the extraction and assess DNA quality. The female-lineage cytochrome c oxidase subunit I (COI) region of the mtDNA was amplified using the COI primers described in Campbell et al. (2005). Amplicon aliquots were stained with SYBR Green and electrophoresed in a 1.5% agarose gel to visualize fragment sizes and confirm successful amplifications. Reactions were purified using exonuclease I and shrimp alkaline phosphatase (EXOSAP). Polymerase-chain-reaction products mixed with EXOSAP were incubated at 37°C for 40 min, followed by 80°C for 20 min to denature any enzymes or remaining primers (as in Hewitt et al. 2019). Samples were Sanger sequenced by Eton Biosciences (etonbio.com) on an Applied Biosystems ABI 3730. The generated sequences were compared to those available on GenBank using BLAST (Altschul et al. 1990). The GenBank sequence with the highest percentage identity score resulting from the BLAST search was chosen as the most likely species and identified as such.

Table 1.

Summary of field-collected and identified Truncilla specimens from rivers in Ontario, Canada.

img-z3-7_99.gif

Traditional morphometric measurements for each specimen were length (maximum distance anterior to posterior) and height (maximum distance dorsal to ventral) measured to the nearest millimeter using Vernier calipers. For field-collected specimens, shell inflation (= width or maximum distance across left and right valves) was also measured. Length-to-height (L/H), length-to-width (L/W), and height-to-width (H/W) ratios were calculated for all field-collected specimens and then arcsine transformed to control for specimen size (as in Inoue et al. 2014).

For geometric morphometric analyses, an image of the left valve of each specimen was uploaded to the MakeFan program of the Integrated Morphometrics Package 8 (IMP8; Sheets 2014). Anchor (type I) landmarks were placed at the tip of the umbo and the point where the hinge ligament ends on the dorsal-posterior side of the valve (landmarks 1 and 2 on Fig. 1). Using MakeFan8, a 40-ray fan was placed at the midpoint between landmarks 1 and 2. Type II landmarks were placed at the point where each ray transected the edge of the shell, with landmark numbers starting at three and consecutively following the rays clockwise around the shell (Fig. 1). We chose to remove the point along the shell margin following landmark 18 because the curve at this point was highly variable among specimens and resulted in placement of the point at different parts of the shell (at the anterior curve of the shell margin, the anterior slope of the hinge, or the anterior of the raised umbo). Type II landmarks 19 and 20 were placed where the rays cross the anterior side of the umbo. The landmarks for each specimen were digitized using MakeFan8. CoordGen8 was used to implement a generalized Procrustes alignment of the digitized landmarks.

Figure 2.

Map of Truncilla donaciformis and T. truncata field-collection site locations and museum specimen localities.

img-z4-1_99.jpg

A principal component analysis (PCA) was applied to the transformed TM ratios (L/H, L/W, H/W) using XLSTAT v. 2018.6 (2018). Discriminant analysis (DA), implemented in XLSTAT using field-identified specimens, was employed to assess the utility of traditional morphometric measurements in species identification. The species memberships of unidentified specimens were predicted using the results of the DA and compared to the confirmed species identifications resulting from the NCBI BLAST search of the COI sequences. An alpha of 0.05 was used in all tests of significance.

CVAGen (in IMP8) was used to implement a canonical variates analysis (CVA, equivalent to the DA used with the TM dataset) to determine if and where significant differences in shell shape exist between the species and to visualize and quantify where the differences were greatest. Differences between shell shapes were visualized using a deformation grid and vectors on landmarks. Museum and field-collected specimens with confirmed identifications (using COI sequences) were used to create the CVA model. The unknown field-collected specimens were used to test the utility of the model using a jackknife assignment test based on the CVA results. We compared the predictions generated by the model to the confirmed species identifications resulting from the NCBI BLAST search of the COI sequences. The species memberships of the unidentified specimens were predicted using a jackknife assignment test based on the CVA results and compared to the confirmed species identifications resulting from the NCBI BLAST search of the COI sequences.

RESULTS

Field collections yielded photographs of 103 individuals: 54 identified as T. truncata, 28 as T. donaciformis, and 21 unidentified Truncilla from the Thames, Welland, Ausable, and Sydenham rivers in Ontario, Canada (Table 1). Museum specimens from the Ohio State University Division of Mollusks and University of Michigan Museum of Zoology were examined and photographed for a total of 44 T. donaciformis and 42 T. truncata specimens (Table 2). Photographs used for morphometric analyses have been submitted to MorphoBank (Project 3457, MorphoBank accession numbers M675689–M675794;  http://morphobank.org/permalink/?P3457).

Table 2.

Summary of Truncilla donaciformis (n = 44) and T. truncata (n = 42) specimens examined, photographed, and digitized at the Ohio State University Division of Mollusks and the University of Michigan Museum of Zoology. Museum collection lot numbers and number of specimens are listed for each species.

img-z5-2_99.gif

DNA Barcoding

Sequencing resulted in COI amplicons averaging 640 bp in length from all 103 individuals, with six unique haplotypes generated for T. donaciformis (GenBank accession nos. MT593033- MT593038) and four unique haplotypes generated for T. truncata (GenBank accession nos. MT594464–MT594467). A comparison of the COI sequences to those on GenBank using NCBI BLAST confirmed the identifications of 35 T. donaciformis and 68 T. truncata specimens. Based on the BLAST search results, one specimen from the Thames River (specimen no. 86; site TR-50) was T. truncata based on its COI sequence but was incorrectly identified in the field as T. donaciformis. Of the 21 unknown Truncilla specimens, 13 were identified as T. truncata, and eight as T. donaciformis.

Traditional Morphometrics

Principal components analysis of the traditional morphometric ratio data showed that the L/H ratio was the primary driver of differences between species shapes (Fig. 3). A t-test of the L/H ratios between species supported the hypothesis that the two groups' values were different (t=–9.81, df = 100, P < 0.001), but there was overlap between the species (Fig. 4). The H/W ratio explained less differentiation between the species, while the L/W ratio was not a diagnostic character between them.

The DA based on the traditional morphometric ratios revealed significant differences between T. truncata and T. donaciformis specimens confirmed with the COI sequence data. The DA correctly identified 93 of 103 specimens when compared to BLAST search results of the COI sequences. The misassigned specimens were three T. truncata and seven T. donaciformis. One of the specimens correctly predicted by the DA was the field-misidentified specimen from the Thames River (specimen no. 86). The DA also was used to assign the unknown Truncilla specimens to a species and was correct for 19 of 21 specimens when compared to BLAST search results of the COI sequences. The two misidentified specimens were a COI-identified T. truncata from the Welland River (site WLR05 specimen no. 10) and a COI-identified T. donaciformis from the Thames River (site TR-24 specimen no. 34). However, the latter specimen had only a 56.1% probability of assignment to the T. truncata group.

Geometric Morphometrics

The CVA revealed differences in shape along a single axis between the COI-confirmed species (Fig. 5). Differences in shell shape along CVA axis 1 were visualized using a deformation grid and landmark vectors (Fig. 6). The main differences were in length near the dorsal margin and in height along the ventral margin of the shell. The jackknife assignment test on the landmark-based CVA dataset resulted in 102 of the 103 field-collected specimens and all of the museum specimens being correctly assigned back to the COI-confirmed or expert-identified (for museum specimens) species. Of the specimens of uncertain species membership collected in the field, 20 of 21 were correctly identified based on their COI sequence. The misidentified specimen, from the Thames River (site TR-50 specimen no. 112), was a COI-identified T. donaciformis that was grouped with T. truncata using geometric morphometrics.

Figure 3.

Principal components analysis (PCA) biplot for arcsine-transformed shell length-to-height (L/H), length-to-width (L/W), and height-to-width (H/W) ratios of field-collected Truncilla specimens. Species identifications were confirmed by comparing COI sequence data to sequences on GenBank.

img-z6-1_99.jpg

DISCUSSION

In identifying putative Truncilla donaciformis and T. truncata specimens, both geometric and traditional morphometrics were shown to be useful, providing a high degree of identification certainty when compared to definitive identifications based on DNA barcodes. Traditional morphometrics, paired with multivariate statistical analyses, were found to correctly identify 90% of the field-identified specimens and 90% of unknown specimens when compared to COI barcode identifications. Multivariate statistical analyses of landmark-based morphometric data provided greater identification accuracy with correct identifications of 99% of field- and museum-identified specimens and 95% of unknown specimens when compared to COI barcode identifications, although in both cases only a single specimen was incorrectly identified.

Shell morphology has long been used to differentiate and identify freshwater mussel species, with varying degrees of reliability (Watters et al. 2009). The utility and accuracy of shell morphology in that capacity has been assessed via mitochondrial DNA barcoding (Hebert et al. 2003; Baird and Sweeney 2011), and recent studies have upheld its merits in assigning individuals to species and groups (Inoue et al. 2013; Guarneri et al. 2014; Inoue et al. 2014; Barreto et al. 2016). However, because freshwater mussel morphology depends, at least partially, on environmental conditions and may vary considerably within species (Hornbach et al. 2010; Suzuki and Nagasawa 2013), care must be taken to systematically quantify the morphological characters being used to identify freshwater mussel specimens. In the sister species T. donaciformis and T. truncata, applying traditional and geometric morphometrics has provided quantifiable means of differentiating between the two. Truncilla truncata is described as being more triangular in shape, while T. donaciformis is more elongate following a continuum in the ratio of shell length to shell height (Watters et al. 2009). Length-to-height ratios >1.5 were 95% likely to be T. donaciformis, and ratios <1.4 were 95% likely to be T. truncata, but there was considerable overlap in the L/H ratio between T. donaciformis and T. truncata specimens. Additionally, the PCA identifies the shell H/W ratio as another important character in differentiating between the species, but this character is difficult to incorporate into field identifications.

Figure 4.

Box plot showing quartiles, means, and outliers of L/H ratios from field-collected Truncilla specimens.

img-z7-1_99.jpg

Both morphometric techniques were able to differentiate between the two species with somewhat different reliability. Traditional morphometric data applied in a discriminant analysis was able to accurately differentiate between the two Truncilla species when compared to DNA barcodes, with an accuracy of 90%. Assignment tests of specimens using the landmark-based geometric morphometric dataset in the CVA provided improved accuracy over traditional morphometrics when compared to the DNA barcodes, with an accuracy of 99% among the specimens analyzed. The analysis of quantifiable morphological characters provided by these two morphometric techniques has the potential to provide researchers and mussel survey teams with greater confidence in assigning specimens to species compared to expert opinion based on qualitative assessment of characters. The general reliability of traditional morphometric techniques in differentiating between T. truncata and T. donaciformis may be sufficient for most specimens and studies, but we recommend using the more reliable geometric morphometric techniques whenever possible, especially for problematic specimens.

Traditional morphometrics using shell measurements, when applied appropriately, have the potential to provide researchers with a great deal more confidence in field identifications than qualitative assessments of shell characteristics (i.e., expert opinion). Conducted using measurements frequently taken in the field—length, width, and height—traditional morphometric data are relatively simple and cost-effective to analyze. The drawback to this ease is accuracy. Based on our findings, 10% of Truncilla specimens identified this way were inaccurately identified. This inaccuracy (i.e., false positives and false negatives) could have some consequences for species conservation efforts in terms of inaccurately estimating the distribution and status of the species, leading to misspent or unnecessary use of resources (Shea et al. 2011).

In contrast to traditional morphometric analyses, landmark-based geometric morphometrics provide a more robust dataset and yield greater accuracy in species identifications (Rohlf and Marcus 1993; Adams et al. 2004; Inoue et al. 2014). However, there is a cost for this accuracy; the methodology requires photos of each specimen, the digitization of landmarks onto those photos, the careful formatting of data, and the use of multivariate statistical analyses. These requirements add a great deal of time and, consequently, higher costs. However, those costs (in time and resources) might be negated by its increased reliability over traditional morphometric analyses; incorrect species identifications made by the latter could result in a waste of time and resources on populations that were erroneously assumed to exist. One relevant example of the implications of misidentification is in the Canadian status assessment of T. donaciformis (COSEWIC 2008), which cites historical records of T. donaciformis from Lyons Creek in the Welland River drainage, as well as a historical record from the Niagara River near the mouth of the Welland. Some of the specimens collected for the present study were taken from other locales in the Welland River and identified as possible T. donaciformis, but mtDNA barcoding and morphometric data both confirmed these as T. truncata. Without an examination of the shell specimens, our findings call into question whether T. donaciformis was ever really present in the Welland River drainage.

Perhaps the only way to identify a freshwater mussel specimen with near perfect certainty is with genetic techniques, but to do so for every specimen collected in a survey is costly and time-consuming, and it requires finding living animals. DNA barcoding techniques (Hebert et al. 2003) using COI and other mtDNA sequences have proven to be useful for identifying most unionoid species (Campbell et al. 2008; Boyer et al. 2011; Inoue et al. 2013, 2014; Keogh and Simons 2019), but there are some exceptions (i.e., within the genera Elliptio and Pleurobema; Inoue et al. 2018). Once a morphological database of specimens with confirmed identifications (i.e., using DNA barcodes) is established, geometric morphometric techniques have the potential to provide a relatively cost-effective and less technology-intensive alternative to DNA barcoding, with a high degree of accuracy that could be used with less reliance on taxonomic/genetic experts.

Figure 5.

Canonical variates analysis biplot of Truncilla specimen shell shape using 20 landmarks. Species identifications were confirmed by comparing COI sequence data to sequences on GenBank or the identification of the museum specimens. The only significant differences in shape between the groups were found along CVA axis 1.

img-z8-1_99.jpg

Misidentifying endangered T. donaciformis in Canada could have negative consequences for efforts directed at conservation of this endangered species, such as in the Welland River (COSEWIC 2008). We used traditional and geometric morphometric techniques, along with COI mtDNA sequencing, to confirm the presence of T. donaciformis in select Canadian watersheds presumed to be inhabited by the species. We can confirm definitively the presence of T. donaciformis in southwestern Ontario in the Thames River, but none of the specimens examined from the Ausable (n=3), Sydenham (n = 1), or Welland rivers were found to be T. donaciformis using DNA barcoding or morphometric analyses. However, our sampling was by no means exhaustive, and T. donaciformis still may be present in these systems. No Truncilla specimens from the lower Grand River in Ontario, where T. donaciformis has recently been found alive (COSEWIC 2008), were collected in the field for this study, but Grand River shell specimens from museum collections (Table 2) did group with DNA barcode-confirmed specimens using landmark-based morphometric analyses, confirming that it was present historically. Museum specimens from Lake Erie grouped with the DNA barcode-confirmed specimens and verify the historical presence of T. donaciformis there.

In addition to definitively identifying unionids collected during field surveys in Ontario, our findings could prove useful for further investigations of T. donaciformis throughout its distribution. This study provides an accurate morphometric framework and dataset to be used in future surveys; our dataset is publicly available (DRYAD Link:  https://doi.org/10.5061/dryad.rn8pk0p6m). Using a similar combination of morphometric and DNA barcoding techniques also could be useful for differentiating among other morphologically similar species. Examples include species in the notoriously difficult-to-identify genera Pleurobema and Fusconaia (Inoue et al. 2018), Pyganodon (Cyr et al. 2007), western North American Anodonta (Chong et al. 2008), and some members of the Lampsilini (Keogh and Simons 2019).

Figure 6.

Deformation grid and vectors showing differences in Procrustes-superimposed shell shape along CVA axis 1 (Fig. 5). The differences in shape are denoted by the direction and relative magnitude of vectors and in the areas where the grid is most deformed from a square. The circles represent the mean shape of T. donaciformis specimens and the ends of the vectors represent the mean shape of T. truncata specimens.

img-z9-1_99.jpg

ACKNOWLEDGMENTS

Funding for this study came from Fisheries and Oceans Canada (DFO) and a Summer Research Scholarship from Central Michigan University (CMU) to TWB. Field specimens were collected by R. Adams, N. Bohlender, A. Boyko, E. Robson, and M. Sheldon (DFO), V. McKay from the Lower Thames Valley Conservation Authority, and D. Jacobs from Meesingw Inc. Thanks to Taehwan Lee (University of Michigan Museum of Zoology) and the late Tom Watters (Ohio State University Museum of Biological Diversity) for facilitating our visits and allowing us to access the collections and photography equipment. Thanks to Jamie Bucholz and Nichelle VanTassel, CMU Biology Department, for their assistance in digitizing landmarks on shell specimens and collecting the DNA barcode data. Two anonymous reviewers and Dr. David Berg provided valuable feedback that improved the paper. Specimens were collected for this study under Fisheries and Oceans Canada Species at Risk Act permit 17-PCAA-00009. This manuscript is contribution no. 142 of the Central Michigan University Institute for Great Lakes Research.

LITERATURE CITED

1.

Adams, D. C., F. J. Rohlf, and D. E. Slice. 2004. Geometric morphometrics: Ten years of progress following the ‘revolution.’ Italian Journal of Zoology 71:5–16. Google Scholar

2.

Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. Journal of Molecular Biology 215:403–410. Google Scholar

3.

Baird, D. J., and B. W. Sweeney. 2011. Applying DNA barcoding in benthology: The state of the science. Journal of the North American Benthological Society 30:122–124. Google Scholar

4.

Barreto, S. B., L. A. Nunes, A. T. da Silva, R. Juca-Chagas, D. Diniz, I. Sampaio, H. Schneider, and P. R. A. de Mello Affonso. 2016. Is Nematocharax (Actinopterygii, Characiformes) a monotypic fish genus? Genome 59:851–865. Google Scholar

5.

Boyer, S. L., A. A. Howe, N. W. Juergens, and M. C. Hove. 2011. A DNA-barcoding approach to identifying juvenile freshwater mussels (Bivalvia: Unionidae) recovered from naturally infested fishes. Freshwater Science 30:182–194. Google Scholar

6.

Burlakova, L. E., D. Campbell, and A. Y. Karateyev. 2019. Status of rare endemic species: Molecular phylogeny, distribution and conservation of freshwater molluscs Truncilla macrodon and Truncilla cognata in Texas. Malacologia 62:345–363. Google Scholar

7.

Campbell, D. C., P. D. Johnson, J. D. Williams, A. K. Rindsberg, J. M. Serb, K. K. Small, and C. Lydeard. 2008. Identification of ‘extinct’ freshwater mussel species using DNA barcoding. Molecular Ecology Resources 8:711–724. Google Scholar

8.

Campbell, D. C., J. M. Serb, J. E. Buhay, K. J. Roe, R. L. Minton, and C. Lydeard. 2005. Phylogeny of North American amblemines (Bivalvia, Unionoida): Prodigious polyphyly proves pervasive across genera. Invertebrate Biology 124:131–164. Google Scholar

9.

Chong, J. P., J. C. Brim-Box, J. K. Howard, D. Wolf, T. L. Myers, and K. E. Mock. 2008. Three deeply divided lineages of the freshwater mussel genus Anodonta in western North America. Conservation Genetics 9:1303–1309. Google Scholar

10.

COSEWIC. 2008. COSEWIC assessment and status report on the Fawnsfoot Truncilla donaciformis in Canada. Committee on the Status of Endangered Wildlife in Canada. Ottawa. vii+ 39 pp. Available at  https://www.registrelep-sararegistry.gc.ca/virtual_sara/files/cosewic/sr_fawnsfoot_0808_e.pdf (accessed April 21, 2020). Google Scholar

11.

Cyr, F., A. Paquet, A. L. Martel, and B. Angers. 2007. Cryptic lineages and hybridization in freshwater mussels of the genus Pyganodon (Unionidae) in northeastern North America. Canadian Journal of Zoology 85:1216–1227. Google Scholar

12.

Folmer, O., M. Black, W. R. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3:294–299. Google Scholar

13.

Guarneri, I., O. P. Popa, L. Gola, L. Kamburska, R. Lauceri, M. Lopes-Lima, L. O. Popa, and N. Riccardi. 2014. A morphometric and genetic comparison of Sinanodonta woodiana (Lea, 1834) populations: Does shape really matter? Aquatic Invasions 9:183–194. Google Scholar

14.

Haag, W. R. 2012. North American freshwater mussels: Natural history, ecology, and conservation. Cambridge University Press, New York. 528 pp. Google Scholar

15.

Hebert, P. D. N., A. Cywinska, S. L. Ball, and J. R. deWaard. 2003. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London Series B: Biological Sciences 270:313–321. Google Scholar

16.

Hewitt, T. L., D. A. Woolnough, and D. T. Zanatta. 2019. Population genetic analyses of Lampsilis cardium (Bivalvia: Unionidae) reveal multiple post-glacial colonization routes into the Great Lakes drainage. American Malacological Bulletin 37:21–34. Google Scholar

17.

Hornbach, D. J., V. J. Kurth, and M. C. Hove. 2010. Variation in freshwater mussel shell sculpture and shape along a river gradient. American Midland Naturalist 164:22–36. Google Scholar

18.

Inoue, K., D. M. Hayes, J. L. Harris, and A. D. Christian. 2013. Phylogenetic and morphometric analyses reveal ecophenotypic plasticity in freshwater mussels Obovaria jacksoniana and Villosa arkansasensis (Bivalvia: Unionidae). Ecology and Evolution 3:2670–2683. Google Scholar

19.

Inoue, K., D. M. Hayes, J. L. Harris, N. A. Johnson, C. L. Morrison, M. S. Eackles, T. L. King, J. W. Jones, E. M. Hallerman, A. D. Christian, and C. R. Randklev. 2018. The Pleurobemini (Bivalvia: Unionida) revisited: Molecular species delineation using a mitochondrial DNA gene reveals multiple conspecifics and undescribed species. Invertebrate Systematics 32:689–702. Google Scholar

20.

Inoue, K., A. L. McQueen, J. L. Harris, and D. J. Berg. 2014. Molecular phylogenetics and morphological variation reveal recent speciation in freshwater mussels of the genera Arcidens and Arkansia (Bivalvia: Unionidae). Biological Journal of the Linnean Society 112:535–545. Google Scholar

21.

Keogh, S. M., and A. M. Simons. 2019. Molecules and morphology reveal ‘new’ widespread North American freshwater mussel species (Bivalvia: Unionidae). Molecular Phylogenetics and Evolution 138:182–192. Google Scholar

22.

NatureServe. 2020. NatureServe Explorer: Fawnsfoot, Truncilla donaciformis. Available at  https://explorer.natureserve.org/Taxon/ELEMENT_GLOBAL.2.121170/Truncilla_donaciformis#Top (accessed April 21, 2020). Google Scholar

23.

Rohlf, F. J., and L. F. Marcus. 1993. A revolution in morphometrics. Trends in Ecology and Evolution 8:129–132. Google Scholar

24.

Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 2344 pp. Google Scholar

25.

Shea, C. P., J. T. Peterson, J. M. Wisniewski, and N. A. Johnson. 2011. Misidentification of freshwater mussel species (Bivalvia:Unionidae): contributing factors, management implications, and potential solutions. Freshwater Science 30:446–458. Google Scholar

26.

Sheets, H. D. 2014. Integrated Morphometrics Package (IMP) version 8. Available at  https://www.animal-behaviour.de/imp/ (accessed April 21, 2020). Google Scholar

27.

Suzuki, M., and H. Nagasawa. 2013. Mollusk shell structures and their formation mechanism. Canadian Journal of Zoology 91:349–366. Google Scholar

28.

Watters, G. T., M. A. Hoggarth, and D. H. Stansbery. 2009. The freshwater mussels of Ohio. Ohio State University Press, Columbus, OH. 400 pp. Google Scholar

29.

Webster, M., and H. D. Sheets. 2010. A practical introduction to landmark-based geometric morphometrics. Pages 163–188 in J. Alroy and G. Hunt, editors. Quantitative methods in paleontology. Vol. 16. Yale University Press, New Haven, CT. Google Scholar

30.

XLSTAT. 2018. Data analysis and statistical solutions for Microsoft Excel. Addinsoft, Paris, France.  https://www.xlstat.com/en/  Google Scholar

31.

Zelditch, M., D. Swiderski, and H. Sheets. 2012. Geometric morphometrics for biologists. 2nd ed. Academic Press, Cambridge, MA. 488 pp. Google Scholar
© Freshwater Mollusk Conservation Society 2020
Tyler W. Beyett, Kelly McNichols-O'Rourke, Todd J. Morris, and David T. Zanatta "Use of Morphometric Analyses and DNA Barcoding to Distinguish Truncilla donaciformis and Truncilla truncata (Bivalvia: Unionidae)," Freshwater Mollusk Biology and Conservation 23(2), 99-108, (28 October 2020). https://doi.org/10.31931/fmbc.v23i2.2020.99-108
Published: 28 October 2020
JOURNAL ARTICLE
10 PAGES


Share
SHARE
KEYWORDS
DNA barcoding
Laurentian Great Lakes drainage
morphometric analysis
shell shape
species identification
RIGHTS & PERMISSIONS
Get copyright permission
Back to Top