University of Texas at Tyler Scholar Works at UT Tyler
Biology Theses Biology
Spring 4-7-2017 Phylogenetic and ecological analysis of two populations of the Kitsatchie Painted Crayfish, Orconectes maletae (Decapoda: Cambaridae) Larrimy Beth Brown University of Texas at Tyler
Follow this and additional works at: https://scholarworks.uttyler.edu/biology_grad Part of the Evolution Commons, and the Population Biology Commons
Recommended Citation Brown, Larrimy Beth, "Phylogenetic and ecological analysis of two populations of the Kitsatchie Painted Crayfish, Orconectes maletae (Decapoda: Cambaridae)" (2017). Biology Theses. Paper 44. http://hdl.handle.net/10950/576
This Thesis is brought to you for free and open access by the Biology at Scholar Works at UT Tyler. It has been accepted for inclusion in Biology Theses by an authorized administrator of Scholar Works at UT Tyler. For more information, please contact [email protected]. Phylogenetic and ecological analysis of two populations of the
Kitsatchie Painted Crayfish, Orconectes maletae (Decapoda:
Cambaridae)
by
Larrimy B. Brown
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Biology
John S. Placyk Jr., Ph.D., Committee Chair College of Arts and Sciences
University of Texas at Tyler April 2017
The University of Texas at Tyler Tyler, Texas
This is the certify that the Master’s Thesis of
Larrimy B. Brown has been approved for the thesis requirement on April 7, 2017 for the Master of Science degree
Acknowledgments
I would like to start by thanking my wonderful and loving husband for supporting me in every way through my journey into biology, as well as, my father and Aunt Robin for taking my phone calls every time I called them frustrated or bustling with excitement. No matter what the situation has been, they have always been there for me through all of this. I would also like to thank my twin sons, Parker and Colton, for being my inspiration and motivation throughout this process. Thank you to my committee members, Lance Williams and Josh Banta, who have offered me so much advice and guidance throughout this project. A special thanks to Marsha
Williams and Kate Hertweck for taking time out of their busy schedules to teach me how to work with many of the statistical and other software packages. Thank you to my advisor, John Placyk, for giving me the opportunity to learn from him and offering me guidance throughout my time at the University of Texas at Tyler. I would like to thank my funding sources, Texas Parks and
Wildlife and the Texas Academy of Sciences. Thank you to Brianna Ciara, Cynthia Ontiveros,
Brenda Arras, and Edith Plants-Paris for all of the hard work they put into helping me with my field collections. Finally, I would like to thank Justin Hernandez for his contributions to the ecological niche modeling portion of this paper.
Table of Contents
List of Tables……………………………………………………………………………………….i
List of Figures………………………………………………………………………………………ii
Abstract……………………………………………………………………………………………...iii
Chapter One: Introduction…………………………………………………………………………..1
Molecular Ecology Introduction..………………………………………….3
Ecological Niche Modeling Introduction..…………………………………5
Chapter Two: Materials and Methods………………………………………………………………6
2014 Texas Survey Methods……………………………………………….6
Molecular Ecology Methods……………………………………………….6
Ecological Niche Modeling Methods………………………………………7
Chapter 3: Results…………………………………………………………………………………12
2014 Texas Survey Results………………………………………………….12
Molecular Ecology Results………………………………………………….15
Ecological Niche Modeling Results…………………………………………21
Chapter 4: Discussion………………………………………………………………………………49
2014 Texas Survey Discussion……………………………………………….49
Molecular Ecology Discussion………………………………………………..49
Ecological Niche Modeling Discussion………………………………………51
Chapter 5: Conclusion………………………………………………………………………….……54
Literature Cited………………………………………………………………………………………56
i
List of Tables
Table 1. Presence points used in all Maxent models…………………………………………………...…15 Table 2. Shows the environmental layers used, the source and website from where the layers were obtained, the scale, and time period…………………………………………………………………..…...15 Table 3. Tissue sample collection information, including: collection site number, county, GPS coordinates, date of collection, and tissue sample names…………………………………….…..….……18
Table 4. Percent sequence divergence of all three genes, comparing the Texas populations and Louisiana populations both within and between populations………………………………………………………...21 Table 5. AMOVA results obtained from the COI gene segment…………………………………………21 Table 6. Pairwise Fst obtained from the COI AMOVA…………………………………………………..22 Table 7. Matrix of significant P-values obtained from the COI AMOVA………………………………..22 Table 8. AMOVA results obtained from the 16s gene segment………………………………………….23 Table 9. Pairwise Fst obtained from the 16s AMOVA…………………………………………………...23 Table 10. Matrix of significant P-values obtained from the 16s AMOVA……………………………….24 Table 11. AUC values obtained from all 3 niche models used in this study……………………………...27 Table 12. Estimates of relative contributions of the environmental variables to the Texas as an individual species Maxent model……………………………………………………………………………………..29 Table 13. Estimates of relative contributions of the environmental variables to the Louisiana as an individual species Maxent model………………………………………………………………………….38
ii
List of Figures
Figure 1. All 25 historical sites sampled in Cypress Creek in East Texas. Samples were not obtained in 14 of the 25 sites sampled (blue). Samples were collected in 11 of the 25 sites (red)……………………17 Figure 2. Haplotype network comparing the unique haplotypes of the Texas and Louisiana subpopulations…………………………………………………………………………………………….24 Figure 3. Habitat suitability map highlighting the areas for each population with the highest habitat suitability for each population of O. maletae with niche overlap…………………………………………28 Figure 4. The logistic response to mean temperature of the driest quarter by the Texas population……..30 Figure 5. The logistic response to stream order by the Texas population………………………………...31 Figure 6. The logistic response to precipitation during the driest quarter by the Texas population……...32 Figure 7. The logistic response to precipitation during the wettest month by the Texas population……..33 Figure 8. Shows the logistic response to soil type (gSSURGO) by the Texas population……………….34 Figure 9. The logistic response to USGS Geology by the Texas subpopulation…………………………35 Figure 10. The logistic response to cumulative square drainage area by the Texas subpopulation………36 Figure 11. The logistic response to the mean temperature during the wettest quarter by the Texas subpopulation……………………………………………………………………………………………...37 Figure 12. The logistic response to the mean temperature during the wettest quarter by the Louisiana subpopulation……………………………………………………………………………………………...39 Figure 13. The logistic response to soil type (gSSURGO) by the Louisiana subpopulation……………..40 Figure 14. The logistic response to cumulative square drainage area by the Louisiana subpopulation….41 Figure 15. The logistic response to stream order by the Louisiana subpopulation……………………….42 Figure 16. The logistic response USGS Geology by the Louisiana subpopulation………………………43 Figure 17. The logistic response to precipitation during the driest quarter by the Louisiana subpopulation……………………………………………………………………………………………..44 Figure 18. The logistic response to precipitation during the wettest month by the Louisiana subpopulation……………………………………………………………………………………………..45 Figure 19. The logistic response to the mean temperature during the driest quarter by the Louisiana subpopulation……………………………………………………………………………………………..46 Figure 20. Test gain for the model run with Texas-Louisiana as a combined species…………………..47 Figure 21. Probability distribution map from the model run with Texas-Louisiana as a combined species…………………………………………………………………………………………………….48 Figure 22. Test gain for the model run with Texas as an individual species…………………………….49 Figure 23. Probability distribution map from the model run with Texas as an individual species………50 Figure 24. Probability distribution map from the model run with Louisiana as an individual species…..51 Figure 25. Test gain for the model run with Louisiana as an individual species…………………………52 iii
Abstract
Phylogenetic and ecological analysis of two populations of the Kitsatchie Painted Crayfish, Orconectes maletae (Decapoda: Cambaridae)
Larrimy Brown
Thesis Chair: John S. Placyk, Jr., Ph.D.
The University of Texas at Tyler May 2017
The primary aim of this study was to assess the conservation status of the East Texas population of the Kistachie Painted Crayfish, Orconectes maletae. Orconectes maletae has been given a
Global Heritage Ranking of G2, or imperiled, because of population declines in the Kisatchie
Bayou drainage in Louisiana; however, similar data for Texas is lacking. For the current study, surveys of 25 historical sites in Texas were conducted in the Cypress Creek and Caddo Lake drainage from June to August 2014 and revealed that O. maletae was either absent or in low enough abundance to evade detection at close to 60% of these sites.
Because of the lack of natural history data available for O. maletae, very little is known about the optimal habitat requirements and the population biology of this organism. The current distribution of the species in Texas and Louisiana is disjunct. As such, the current study used molecular genetic techniques and ecological niche modeling to determine if the geographically isolated subpopulations of O. maletae in East Texas and Louisiana are differentiated genetically or ecologically.
Molecular analyses were conducted using the mitochondrial gene segments COI and 16S as well as the nuclear gene segment GADPH. Results of the AMOVAs for COI and 16S suggests
I iv
that significant divergence (possibly even speciation) may have occurred between the two subpopulations. The results of the percent sequence divergence, pairwise sequence divergence, and fixation indices all supported this conclusion. The nuclear gene AMOVA did not yield significant results. A haplotype network was generated to compare the unique haplotypes from both subpopulations and showed separation of the Texas and Louisiana haplotypes into unconnected networks because they could not be linked with 95% certainty or greater.
The ecological niche modeling data obtained for this study further supported the idea that divergence may have occurred, as the models proved to be more fit when run as separate species than when they were run together as a single species. Specifically, the Texas Kisatchie Painted
Crayfish differ in habitat associations from their counterpart in Louisiana indicating that they likely occupy separate niches in their respective states.
The molecular and ecological data together support the notion that the subpopulations have undergone divergence resulting in the emergence of a cryptic species that, superficially, looks morphologically identical to its counterpart. Because of the evidence supporting divergence, and the decline in optimal habitat leading to the absence of O. maletae in both East Texas and
Louisiana, it might be appropriate to consider federal conservation for these organisms. Action must be taken to preserve the remaining habitat in both states to conserve these organisms and prevent one or both from becoming extinct. However, before this claim can be made, further analysis of these organisms should be conducted.
In future studies, more samples should be collected from both states for further phylogenetic analysis, where a phylogeny is constructed to enhance the understanding of the
v
lineage of O. maletae in both Texas and Louisiana. It would also be appropriate to ground truth the Maxent models generated for this analysis. This would both verify if the areas predicted to be suitable for O. maletae support populations, and possibly allow for range expansion if the organism is found to exist outside of its known range.
vi
Chapter 1: Introduction Freshwater ecosystems are faced with many environmental challenges that have led to the widespread loss of biodiversity. The major threat to freshwater ecosystems is habitat loss
(Wilcove 1998), followed by pollution (Lake 2000) and introduction of foreign species
(Mathews 2008). The loss of biodiversity in these aquatic systems is especially prevalent among freshwater crayfish. Crayfish play a vital role in freshwater ecosystems as a major component at multiple trophic levels. They are a major food source for fish and reptile predators (Garvey
1994), and they are omnivores that feed on macroinvertebrates and macrophytes (Lodge 1994).
Crayfish also play a substantial role in lower trophic levels as decomposers of detritus (Usio
2000). In North America, there are over 360 described species of crayfish (Crandall 2008).
Current estimates predict that up to 48% of North American crayfish species are threatened, and one third of all crayfish species worldwide are threatened (Richman 2015).
For one such species, the Kisatchie Painted Crayfish, Orconectes maletae, there is insufficient data to determine its conservation status in Texas. This lack of data is because of its relative rarity, as this species is known to occur only in two locations: the Cypress Creek drainage in East Texas and the Kisatchie Bayou drainage in Louisiana (Fitzpatric 1987, Walls
2005). In Louisiana, Walls (2009) indicated that O. maleate populations are declining with recent survey data indicating that they are now locally extinct from four of seven historic collection sites. Degradation of the Louisiana habitat has been observed, with excessive sedimentation in much of the species range (Walls 2009). As a result, O. maletae has been given the Global Heritage Rank Status of G2, or imperiled, by NatureServe (Taylor 2007, NatureServe
2009) and the status of Threatened by the American Fisheries Society (Taylor 2007). However,
- 1 -
in the state of Louisiana, O. maletae is currently listed as “data deficient” because of a poor knowledge of the species (IUCN Red List of Threatened Species 2010).
The Endangered Species Act deems that conservation efforts must be made to protect both imperiled species and the ecosystems on which these organisms depend (ESA, 16 U.S.C.
1531-1544). As such, O. maletate is being considered for conservation status listing in Texas and
Louisiana. However, because the species as a whole has been deemed data deficient (IUCN Red
List of Threatened Species 2010), it would be optimal to collect data on both Texas and
Louisiana populations. It is also important to determine if the two populations are genetically similar enough to classify them as belonging to the same species. This analysis is necessary because North American crayfish have a high rate of diversity, and, as result, we see a higher than normal rate of speciation events in this taxa (Mathews 2008). Furthermore, it is not well understood how phylogeography has played a role in in relation to this high level of diversity
(Scholtz 2002, Fetzner 2003, Buhay 2005). It is vital to understand if the geographically isolated populations of O. maletae are diverging. As the species is only known to occur in two freshwater systems, then the individual populations of this species need to be closely examined at the genetic and ecological level for a complete evaluation of the species for listing and conservation status. This analysis needs to be conducted in both states, as the information is invaluable in making conservation decisions regarding both the organism and its ecosystem. New information could lead to changes in how O. maletae is managed both at the state and federal levels.
While individual O. maletae from Texas and Louisiana appear to be similar in terms of morphology, it is unclear if the two populations have undergone any divergence as a result of the current landmass isolating them and likely preventing gene flow. The aim of this study, therefore, was to determine 1) if O. maletae still occur in each of the historical locations within
- 2 -
the Cypress Creek drainage in East Texas, and 2) if the Texas and Louisiana populations have undergone divergence because of geographic and reproductive isolation. This study used molecular ecology methodologies, ecological niche modeling, historical data, and field collections to analyze both populations of O. maletae.
Molecular Ecology Introduction
Classical taxonomy has knowledge gaps that have necessitated modern reevaluations of standing classification by incorporating molecular genetic techniques. Traditionally, morphology, behavior, geography, and ecology have been used to classify crayfish species. Because molecular techniques have been applied to the evolutionary history of the crayfish superfamily, Astacoidea, many issues in the existing classification have been identified, with many cryptic species being discovered (Sinclair 2004). The inadequacies of taxonomic classification has led to challenges in conservation efforts.
Genetic barcoding is a molecular technique in which a segment of DNA is identified as the standardized gene segment for species identification (Astrin 2006). This short segment of DNA acts as a molecular tag that is unique to a species (Herbert 2003). In crayfish, this gene segment is the mitochondrial cytochrome c oxidase subunit I (COI) (Mathews 2008). This molecular technique is used as a means of re-identification of organisms at the species level, as the standardized genetic information is incorporated into the existing classification system (Astrin
2006). In conjunction with genetic barcoding, other genetic markers are often used to get a better understanding of the phylogenetics of the target organism. In this case the mitochondrial gene 16S and nuclear gene GADPH were used.
- 3 -
Mitochondrial genes are often chosen in phylogenetic studies because of the high rate of mutations that occur in mitochondrial DNA, maternal inheritance, and lack of recombination.
Mitochondria lack mechanisms to repair errors during replication, resulting in a higher rate of evolution; this makes mitochondrial DNA ideal for comparing closely related organisms
(Munasinghe 2003).
Percent sequence divergence is a statistical technique used to calculate the amount of variation of individual nucleotides between segments of DNA. In crayfish, the 16S gene region has been observed to have less variation than the COI gene region. Sequence divergences vary from 0.84 % at the intraspecific level to 7.63% at the interspecific level for 16S. For COI, the sequence divergences varied from 2.42% at the intraspecific level to 15.53% at the interspecific level (Munasinghe 2003).
Fixation indices are the primary way to measure genetic differentiation using analysis of molecular variance (AMOVA) (Bird 2011). Wright’s Fst is a fixation index that measure genetic differentiation on a scale of 0.0 – 1.0, with 0.0 being the least genetically differentiated and 1.0 being the most genetically differentiated (Write 1943).
The AMOVA also used the following fixation indices to measure genetic differentiation:
Fst, Fsc, and Fct. These fixation indices are measured on the same scale as above. Fst measures the amount of genetic differentiation between individuals within a subpopulation, Fsc measures the amount of genetic differentiation between subpopulations, and Fct measures the amount of genetic differentiation between groups of subpopulations (Mariani 2005, Schneider 2000).
- 4 -
Ecological Niche Modeling Introduction
Ecological niche models are used to assess the relationship between predetermined environmental variables, ArcGIS layers, and the known location of a species. A probability distribution model is then generated by projecting the relationship across a landscape. This model shows the habitat suitability for the species based on the layers used. Maximum entropy modeling
(Maxent) uses presence only data to find the probability distribution of maximum entropy, with the given constraints of known locations and environmental variables (Raxworthy 2007).
Maxent models are invaluable in conservation efforts, as they can lend insight into where rare organisms are most likely to be found (Phillips 2008). These models can also lend insight into which areas would be most suitable for reintroduction of endangered organisms based on habitat suitability (Ghia 2013).
O. maleate has been observed in both clear and muddy water. They have been documented to prefer clear, flowing, sandy bottomed rivers. They can be found under debris, rocks, and along riverbanks in vegetation (Walls 2009). The ecological niche models generated for this study will lend insight into both the Louisiana and Texas subpopulations, and will test if they occupy the same niche as inferred from habitat associations.
- 5 -
Chapter 2: Materials and Methods
2014 Texas Survey Methods
We surveyed all 25 of the historic O. maletae sites in the East Texas Cypress Creek system to determine if the species is in decline (not found at all 25 historic sites) or stable (still found at all 25 historic sites). Specimens were also collected from one of the historic Louisiana sites, a tributary of the Kisatchie Bayou (Walls 2005).
Specimens were collected from each site using minnow traps baited with canned meat in the form of sardines and hot dogs (Hobbs 2010), as well as through dip netting along the banks and in riffles (Rabeni 1997). Each site was visited a minimum of two times, which allowed for two rounds of dip netting and traps being set for 24-48 hours. From each site, one specimen was sacrificed as the representative of the species, and preserved in 75% ethanol (Crandall 2007). All other collected specimens had tissue collected in the form of one, or both, of the 5th pereiopod(s), preserved in 95% ethanol and then frozen for genetic analyses (Mathews 2008).
Molecular Ecology Methods
DNA was extracted from tissue samples using the GE Healthcare illustra tissues and cells genomicPrep Mini Spin Kit. The genes examined to determine relatedness between Texas and
Louisiana subpopulations were two mitochondrial and one nuclear gene: 16S, COI, and GADPH, respectively. The mtDNA was amplified using the following primers: 16S gene, 16S-1472
(Schubart 2000) and 16S-L2 (Mathews 2002); COI gene, orcoCOIF and orcoCOIR (Mathews
2008). The nuclear gene was amplified using the following primers: GADPH gene, G3PCq157F and G3PCq981R (Buhay 2007). Qiagen TopTaq DNA Polymerase kits were used, for each 20 µL
- 6 -
PCR reaction: 0.1 µL Taq, 7.1 µL H2O, 2.0 µL Coral Load, 2.0 µL Q, 0.4 µL DNTPs, 2.0 µL x10 buffer, 1.0 µL each primer, and 2.4 µL DNA was used. The polymerase chain reaction (PCR) conditions for each of the amplified genes consisted of: denaturation for 2 mins at 95C, 40 cycles of 30 sec at 95C, 30 sec at the following annealing temperatures: 16S: 48C; COI: 52C; GAPDH:
60C, and extension for 10 min at 72C (Mathews 2008).
PCR products were purified using the Omega bio-tek Cycle Pure Kit. The purified PCR products were then concentrated and shipped to Eurofins MWG Operon for sequencing using
BigDye® Terminator v 3.1 Cycle Sequencing kits (Applied Biosystems). The genes were sequenced on an automatic sequencer at Eurofins MWG Operon. Then the genes were manually edited using Sequencher (Bromberg 1995), and aligned using ClustalX (Thompson 2002). Final editing was conducted using Mesquite (Maddison 2001).
AMOVA analysis was conducted using Arlequin version 3.5.2.2. This program computed the conventional F-statistics from haplotype frequencies (Weir 1984, Excoffier 1992, and Weir
1996). Percent sequence divergence was calculated manually by counting the polymorphic sites and dividing that number by the total number of base pairs in the sequence (Henikoff 1994, Wetzer
2001). A haplotype network for the COI gene was generated using the program TCS 1.21. Each haplotype represents a unique gene sequence in the COI gene segment used in this study.
Ecological Niche Modeling Methods
Three regions were chosen for this study: Arkansas Red-White 11, Texas 12, and Lower
Mississippi 08. This includes the portions of the Red River and its drainages and tributaries. These regions incorporate all of O. maletae’s known habitat, as well as, regions of river which could potentially be suitable habitat.
- 7 -
Historical site data for the Texas subpopulation were accessed through the Texas Parks and
Wildlife data base in May 2014 (http://tpwd.texas.gov/huntwild/wild/wildlife_diversity/txndd/).
The historical site data for the Louisiana subpopulations were obtained from a report written over the crayfish of the Kisatchie National Forest (Walls 2005). The presence points for Texas were from the 2014 survey (Table 1). The presence points for Louisiana were from the historical sites reported to have O. maletae present in the Kisatchie National Forest (Walls 2005) and from field collections conducted in summer 2014 and 2015 (Table 1).
Maximum entropy modeling methods (Maxent) were applied using Maxent version 3.3.3k
(Phillips 2006; http://cs.prinston.edu/~schapire/Maxent). To generate the models, Texas and
Louisiana presence points were used from field collections and the reported presence points found in the report of Kisatchie National Forest (Walls 2005). Presence points, as well all data used in this analysis, were projected to North American Datum (NAD) 1983 UTM zone 14N in ArcGIS
10.3 (ESRI 2014).
A map of habitat suitability with niche overlap from the Maxent model predictions for O. maleate was generated using ArcGIS 10.3. This map showed the areas in both Texas and Louisiana that were suitable for both populations of O. maleate individually, and what habitat might be suitable for both populations (Carter 2011).
A test of spatial autocorrelation for these layers was run using ENMtools 1.4.4. Spatially correlated layers were removed (a total of 6 layers out of 19), and a series of models were run. One model was generated using the presence data from Texas and Louisiana formatted with both subpopulations combined into a single species, and separate models were generated with the data formatted as separate species, with each state’s subpopulation representing a single species. The
- 8 -
layers retained for the Maxent analysis were as follows: stream order, soil properties, mean temperature driest quarter, mean temperature wettest quarter, maximum temperature warmest month, precipitation wettest month, precipitation driest quarter, precipitation wettest quarter, isothermality, annual velocity (measured in feet per second), USGS geology, cumulative square drainage area (measured in miles squared), and ESRI landcover (Environmental Systems Research
Institute) (Table 2).
A cross-validation approach (Pearson et al. 2007) was used to subdivide the datasets into the training data (species occurrence locations used to develop models) and test data (species occurrence locations that were not used in model development). A “leave-one-out” or “n-1” cross- validation method, as previously described by Pearson et al. (2007), was used so that the number of folds for each species was equal to the number of samples, with each fold containing n -1 observations, where n is the total sample size. Each fold had a single test data point, and each presence location was the test data point, in turn, for a separate fold. Model statistics were averaged across the n folds for each species.
All environmental layers were converted to raster format with a 90 x 90-meter cell size that was clipped to study’s extent. Maxent settings were set to default, except for random seed and a threshold set for minimizing training presence. Because no absence points are used when generating a Maxent model, 10,000 psuedoabsence points are generated automatically (Phillips
2005).
Area under the curve (AUC) measures the probability that presence points will have a higher habitat suitability score than the randomly generated pseudoabsence points (Phillips 2008). AUCs with a value of higher than 0.75 are considered useable. The higher the AUC, the better the
- 9 -
predictive capability of the model (Elith 2002). The test AUCs represent the average percentage of the pseudoabsence data with lower habitat suitability scores than single “test” presence locations left out of the model building process for each model fold. Importantly, this model validation procedure is based on data points (test data) that were naïve to the model building process for each model fold, and thus represent a form a ground-truthing of the models with independent data.
To quantify the relative importance of the individual environmental variables to the models, the fit of each full model was compared to reduced univariate models (Phillips 2006). If an environmental variable accounted for most of the model fit when modeled by itself (as compared to the full model that was based on all the environmental variables), then the environmental variable was considered important in determining the varying habitat suitability of the landscape for that model (Phillips 2006). Model fit was measured with the gain statistic (Phillips 2006). Gain is a likelihood (deviance) statistic that measures the model performance compared to a model that assigns equal habitat suitabilities to all areas of the landscape. Taking the exponent of the final gain gives the (mean) probability of the presence sample(s) compared to the pseudoabsences. For instance, a gain of 3 means that an average presence location has a habitat suitability of e3 = 20.1 times higher than an average pseudoabsence site. The test gains that were reported are the averages of test gains of the single “test” presence locations left out of the model building process for each model fold.
An average habitat suitability map was generated for the Texas and Louisiana subpopulations as a single species, and then two maps were generated with the two subpopulations divided into separate species. This shows the rivers in a range of colors from dark blue to red, the colors correspond to a scale from 0.0-1.0, with 0.0 (dark blue) being the least habitat suitable for the species and 1.0 (red) being the most suitable habitat. These colors are a visual representation of
- 10 -
the habitat suability as predicted by the model. The habitat suitability maps are generated to estimate distribution across geographic space (Phillips 2006).
Table 1. Presence GPS points used in all Maxent models.
State County Latitude Longitude TX Upshur 32.775922 -94.946365 TX Upshur 32.79811 -95.04985 TX Upshur 32.67281 -94.75155 TX Harrison 32.63586 -94.67286 TX Marion 32.788948 -94.515727 TX Marion 32.7497 -94.49978 TX Marion 32.75633 -94.34306 TX Titus 33.02177 -94.88128 TX Titus 33.073109 -94.96546 TX Franklin 33.056416 -95.14274 LA Natchitoches Parish 31.548072 -93.235596 LA Natchitoches Parish 31.581799 -93.278477 LA Natchitoches Parish 31.643775 -93.234848 LA Natchitoches Parish 31.648233 -93.213169 LA Natchitoches Parish 31.440651 -93.089903 LA Natchitoches Parish 31.779315 -93.347079 LA Natchitoches Parish 31.410108 -93.169194
Table 2. All environmental layers used, the source and website from where the layers were obtained, the scale, and time period.
- 11 -
Chapter 3: Results
2014 Texas Survey Results
A survey of the historic sites of O. maletae was conducted between June – August 2014.
Although not much is known about the mating habits of O. maletae specifically, most temperate crayfish in North America mate in early spring (Berrill 1982). As such, summer months were chosen to reduce the chance of interference with reproductive efforts of the wild crayfish and to avoid catching females holding a clutch of eggs or offspring. Exhaustive trapping (Hobbs 2010), and dip netting (Rabeni 1997) were conducted in all of the sites surveyed. The species of interest was successfully collected from 10 of the 25 sites, but no individuals were detected or obtained at the other 14 sites (Figure 1). One site was inaccessible because it was private property, and the owner could not be contacted to grant permission for access to the land(Table 3).
Because of the extent of the trapping and dip netting in each site, it was inferred that 14 of the sites may no longer support this species. This was a result of the populations in these sites being absent or having reached such low numbers as to evade detection. This indicates that the O. maletae subpopulation may be in decline in East Texas.
- 12 -
Figure 1. All 25 historic sites sampled in Cypress Creek in East Texas. Samples were not obtained in 14 of the 25 sites sampled (blue). Samples were collected in 11 of the 25 sites (red).
- 13 -
Table 3. Tissue sample collection information, including: collection site number, county, GPS coordinates, date of collection, and tissue sample names.
Site County Latitude Longitude Date Collected Tissue Sample Name Number 1 Upshur 32.77541 -94.94578 6/12/2014 OM1, OM2 2 Upshur 32.79811 -95.04985 6/12/2014 OM3, OM4 3 Upshur 32.79075 -95.09458 6/12/2014 None Collected 4 Upshur 32.70313 -94.80828 7/4/2014 None Collected 5 Gregg 32.67281 -94.75155 7/4/2014 OM5 6 Harrison 32.63586 -94.67286 7/4/2014 OM6 7 Harrison 32.62358 -94.57773 7/4/2014 None Collected 8 Harrison 32.62743 -94.51598 7/19/2014 None Collected 9 Harrison 32.67161 -94.42331 7/19/2014 None Collected 10 Marion 32.79878 -94.5897 7/19/2014 None Collected 11 Marion 32.78889 -94.51634 7/28/2014 OM18 12 Marion 32.78569 -94.51417 7/19/2014 None Collected 13 Marion 32.7497 -94.49978 7/25/2014, OM7-OM13, OM14-OM17 7/28/14 14 Marion 32.84863 -94.43212 7/28/2014 None Collected 15 Cass 32.89255 -94.44119 7/25/2014 Site inaccessible 16 Marion 32.75633 -94.34306 8/2/2014 OM19-OM20 17 Marion 32.73613 -94.28851 8/6/2014 None Collected 18 Harrison 32.69605 -94.18785 8/6/2014 None Collected 19 Titus 33.02177 -94.88128 8/8/2014, OM21-39, OM45-50 8/13/14 20 Titus 33.07185 -94.96546 8/8/2014 OM40-44 21 Titus 33.09425 -95.01338 8/8/2014 None Collected 22 Titus 33.04824 -95.09604 8/13/2014 None Collected 23 Franklin 33.0511 -95.14247 8/13/2014 OM51 24 Franklin 33.05247 -95.2206 8/13/2014 None Collected 25 Franklin 33.02401 -95.27005 8/13/2014 None Collected 26 Natchitoches 31.410108 -93.169194 May 2014, June LOM 1-8, LOM9-17 Parish 2015
- 14 -
Molecular Ecology Results
The analysis of the COI gene segment generated 592 bp for 10 O. maletae from Louisiana and 50 O. maletae from Texas (GenBank Accession #’s KY788237-43). Percent sequence divergence for the COI gene was 1.86% within the Texas subpopulations, 0.17% divergence within the Louisiana subpopulations, and a 5.91% between the two groups of subpopulations (LA vs. TX)
(Table 4). Additionally, AMOVA results for the COI sequence showed no variation between subpopulations (Fsc = 0.00), but did show significant variation between individuals of the subpopulations (Fst = 0.92) and between TX and LA (Fct = 0.93) (Schneider 2000) (Table 5).
These data indicate that 92.68% of the variation observed is due to difference between the Texas and Louisiana subpopulations. Only 7.32% of the variation is due to differences within the Texas subpopulation (Table 5).
The pairwise sequence divergence obtained from the COI AMOVA showed divergence occurring between the Texas and Louisiana groups of subpopulations, with values ranging between
0.93-1.00 (Table 6). The matrix of significant Fst P-values obtained from the COI AMOVA showed significant genetic differentiation between 6 of the 10 Texas subpopulations when compared to the Louisiana subpopulation (Table 7).
The analysis of the 16S gene segment generated 552 bp for 10 O. maletae from Louisiana and 38 O. maletae from Texas (GenBank Accession #’s KY788229-33). Percent sequence divergence for the 16S gene was 0.91% within the Texas subpopulations, and 2.37% between the two subpopulations (Table 4). AMOVA results for the 16S sequence showed no variation between the subpopulations (Fsc = 0.00), but significant variation was detected between individuals of the subpopulations (Fst = 0.99) and between TX and LA (Fct = 0.99) (Schneider 2000) (Table 8).
- 15 -
These data indicate that 99.0% of the variation observed is due to difference between the Texas and Louisiana subpopulations. Only 1.0% of the variation is due to differences within the Texas subpopulation (Table 8).
The pairwise sequence divergence obtained from the 16S AMOVA showed divergence occurring between the Texas and Louisiana groups of subpopulations, with values ranging between
0.99-1.00 (Table 9). The matrix of significant Fst P-values obtained from the 16S AMOVA showed significant genetic differentiation between 5 of the 10 Texas subpopulations when compared to the
Louisiana subpopulation (Table 10).
The analysis of the GADPH gene segment generated 816 bp for 8 O. maletae from
Louisiana and 23 O. maletae from Texas (GenBank Accession #’s KY788234-36). Percent sequence divergence was not significant for the GADPH gene segment, between populations or within the populations (Table 4). AMOVA results were also not significant for the GADPH gene segment. This is expected with nuclear DNA as the rate of mutations is significantly less that that seen in mitochondrial DNA (Munasinghe 2003).
The haplotype network generated for this study showed the Texas and Louisiana subpopulations to have unique haplotypes with regards to their individual subpopulations (Figure
2). The haplotypes with the highest probability of being ancestral for each subpopulation is displayed as a square, while the move diverged haplotypes are displayed as ovals. The size of the oval or square corresponds to the number individuals represented by that haplotype. The Texas samples had similar enough haplotypes to share one network with the most ancestral haplotype being represented by the sample TXOM 1.
- 16 -
Table 4. Percent sequence divergence of all three genes, comparing the Texas populations and Louisiana populations.
Gene Texas Population Louisiana Population Combined Populations COI 1.86% 0.17% 5.91% 16s 0.91% 0.18% 2.37% GADP 0.12% 0.12% 0.25% H
Table 5. AMOVA results obtained from the COI gene segment.
Source of sum of Percentage of d.f. Variance components variation squares variation
Among groups 1 210.553 12.64324 Va 92.68 Among populations within groups 9 5.96 0.00000 Vb 0 Within populations 49 54.72 1.11673 Vc 8.19 Total 59 354.204 13.64202
- 17 -
Table 6. Pairwise Fst obtained from the COI AMOVA.
Pairwise difference (Fst) TX 1 TX2 TX3 TX4 TX5 TX6 TX7 TX8 TX9 TX10 LA1 TX1 0.00 ------TX2 0.00 0.00 ------TX3 0.00 0.00 0.00 ------TX4 0.00 0.00 0.00 0.00 ------TX5 0.00 0.00 0.00 0.00 0.00 ------TX6 0.00 0.00 0.00 0.00 0.00 0.00 ------TX7 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ------TX8 0.00 0.00 0.00 0.00 0.00 0.00 0.06 0.00 ------TX9 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -- -- TX10 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -- LA1 0.99 1.00 1.00 1.00 1.00 0.95 0.97 0.93 0.96 1.00 0.00
Table 7. Matrix of significant P-values obtained from the COI AMOVA.
Matrix of significant Fst P values TX 1 TX2 TX3 TX4 TX5 TX6 TX7 TX8 TX9 TX10 LA1 TX1 TX2 - TX3 - - TX4 - - - TX5 - - - - TX6 - - - - -
TX7 ------TX8 ------TX9 ------TX10 ------LA1 + + - - - + + + + -
- 18 -
Table 8. AMOVA results obtained from the 16S gene segment.
sum of Variance Percentage of Source of variation d.f. squares components variation Among groups 1 347.896 21.86283 Va 99.44 Among populations within groups 9 0.626 0.00000 Vb 0 Within populations 38 5.681 0.14951 Vc 0.68 Total 48 354.204 21.98592
Table 9. Pairwise Fst obtained from the 16S AMOVA.
Pairwise difference (Fst) TX 1 TX2 TX3 TX4 TX5 TX6 TX7 TX8 TX9 TX10 LA1 TX1 0.00 TX2 0.00 0.00 TX3 0.00 0.00 0.00 TX4 1.00 1.00 1.00 0.00 TX5 1.00 1.00 1.00 0.00 0.00 TX6 0.00 0.00 0.00 0.00 0.00 0.00 TX7 0.00 0.50 0.00 0.00 0.00 0.00 0.00 TX8 0.00 0.28 0.00 0.00 0.00 0.00 0.02 0.00 TX9 1.00 1.00 1.00 0.00 0.00 0.08 0.00 0.00 0.00 TX10 1.00 1.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 LA1 0.80 0.83 0.80 0.80 0.80 0.49 0.67 0.56 0.80 0.80 0.00
- 19 -
Table 10. Matrix of significant P-values obtained from the 16S AMOVA.
Matrix of significant Fst P values TX 1 TX2 TX3 TX4 TX5 TX6 TX7 TX8 TX9 TX10 LA1 TX1 TX2 -
TX3 - - TX4 - - -
TX5 - - - - TX6 - - - - - TX7 ------
TX8 ------TX9 ------TX10 ------LA1 + + - - - + + + - -
- 20 -
Figure 2. Haplotype network comparing the unique haplotypes of the Texas and Louisiana COI subpopulations.
Ecological Niche Modeling Results
The area under the curve (AUC) of the combined data including all presence points collected from Texas and Louisiana was found to be 0.851 (Table 11). When the models were run as separate species, the AUC was shown to be higher in both models, with Texas having an AUC of 0.958 and Louisiana 0.917 (Table 11). As such, the single species model was abandoned in favor for the separate species model.
The habitat suitability map generated showed no niche overlap between the two populations of O. maletae. Each population had suitable habitat in their respective state that was separated by approximately 150 km of unsuitable habitat for both populations. This model indicated that there was no shared habitat between the two populations of O. maletae (Figure 3).
The Texas subpopulation Maxent model showed the following geographic layers to be the most important predictors of habitat suitability: mean temperature of the driest quarter, stream order, mean precipitation of the driest quarter, and precipitation of the wettest month (Table 12).
The most important environmental layer for this subpopulation was the mean temperature of the driest quarter, with the optimal temperature being 29.61 °C (Figure 4). This was followed by stream order, which showed this subpopulation to be more heavily distributed in larger order streams at 4-5 (Figure 5). Then was mean precipitation of the driest quarter, with the optimal precipitation being 18.68 mm (Figure 6). Finally was precipitation of the wettest month, showing an optimal precipitation of 11.16 mm (Figure 7).
The following layers were not highly important to the model, but can lend insight into the ecological niche of the Texas subpopulation in comparison to the Louisiana subpopulation. These layers were as follows: soil types, USGS geology, cumulative square drainage area, and the mean
- 21 -
temperature of the wettest quarter. The soil types (gSSURGO) important to the Texas subpopulations were Estes clay loam that is frequently flooded, Mooreville-Mantachie loam complex that is frequently flooded, and Gallime fine sandy loam with 5-12% slopes (Figure 8).
The Geology layer for this subpopulation predicted a heavier distribution in streams with sandy bottoms (Figure 9). The cumulative square drainage area showed the optimal drainage to be -0.17 x 105 푘푚2 (Figure 10). The mean temperature of the wettest quarter showed the optimal temperature to be 29.04 °C (Figure 11).
The Louisiana subpopulation Maxent model showed the following geographic layers to be the most important predictors of habitat suitability: mean temperature of the wettest quarter, soil types, cumulative square drainage area, and stream order (Table 13). The most important environmental layer for this subpopulation was the mean temperature of the wettest quarter, with the optimal temperature being 8.94 °C (Figure 12). This was followed by soil types, which showed the soil types important to this subpopulation to be Gutyon silt loam, 5-12% slopes and Guyton-
Lotus association that is frequently flooded (Figure 13). Then was cumulative square drainage area, with the optimal drainage being -0.17 x 105 푘푚2 (Figure 14). Finally there was stream order, which showed this subpopulation being more heavily distributed in smaller order streams, specifically 2 (Figure 15).
The following layers were not highly important to the model, but can lend insight into the ecological niche of the Louisiana subpopulation in comparison to the Texas subpopulation. These layers were as follows: USGS geology, mean precipitation of the driest quarter, precipitation of the wettest month, and mean temperature of the driest quarter. The geology layer for this subpopulation predicted a heavier distribution in streams with clay or mud bottoms (Figure 16).
The mean precipitation of the driest quarter showed the optimal precipitation to be 33.32 mm
- 22 -
(Figure 17). The precipitation of the wettest month showed the optimal precipitation to be 17.97 mm (Figure 18). The mean temperature of the driest quarter showed the optimal temperature to be
29.61 °C (Figure 19).
Table 11. AUC values obtained from all 3 Maxent models used in this study.
AUC Combined Model 0.851
AUC Texas Model 0.958
AUC Louisiana Model 0.917
- 23 -
Figure 3. Habitat suitability map highlighting the areas for each population with the highest habitat suitability for both populations of O. maletae with niche overlap.
- 24 -
Table 12. Estimates of relative contributions of the environmental variables to the Texas as an individual species Maxent model.
Variable Percent contribution Permutation importance Mean temp d.q. 32 74.7 stream order 30 7.3 precip d.q. 21.2 2.1 precip w.m. 14.1 13.7 isotherm 0.8 0.9 geology 0.6 0.1 annual velocity 0.4 0.5 cummulative drainage 0.3 0 GSSURGO 0.3 0.1 Mean Temp W.Q. 0.2 0.4 Precip W.Q. 0.1 0.3 esri landcover 0 0 max temp w.m. 0 0
- 25 -
Figure 4. The logistic response to mean temperature of the driest quarter by the Texas subpopulation.
- 26 -
Figure 5. The logistic response to stream order by the Texas subpopulation.
- 27 -
Figure 6. The logistic response to precipitation during the driest quarter by the Texas subpopulation.
- 28 -
Figure 7. The logistic response to precipitation during the wettest month by the Texas subpopulation.
- 29 -
Figure 8. Shows the logistic response to soil type (gSSURGO) by the Texas subpopulation.
- 30 -
Figure 9. The logistic response to USGS Geology by the Texas subpopulation.
- 31 -
Figure 10. The logistic response to cumulative square drainage area by the Texas subpopulation.
- 32 -
Figure 11. The logistic response to the mean temperature during the wettest quarter by the Texas subpopulation.
- 33 -
Table 13. Estimates of relative contributions of the environmental variables to the Louisiana as an individual species Maxent model.
Variable Percent contribution Permutation importance
Mean temp w.q. 82.3 88.6 gssurgo 11.4 4.7 cummulative drainage 2.4 0.2 Stream order 1.7 0.5 isotherm 1 0.5 max temp w.m. 0.6 1.6 geology 0.5 2 mean temp d.q. 0 1.9 precip w.q. 0 0 precip w.m. 0 0 precip d.q. 0 0 esri landcover 0 0 annual velocity 0 0
- 34 -
Figure 12. The logistic response to the mean temperature during the wettest quarter by the Louisiana subpopulation.
- 35 -
Figure 13. The logistic response to soil type (gSSURGO) by the Louisiana subpopulation.
- 36 -
Figure 14. The logistic response to cumulative square drainage area by the Louisiana subpopulation.
- 37 -
Figure 15. The logistic response to stream order by the Louisiana subpopulation.
- 38 -
Figure 16. The logistic response USGS Geology by the Louisiana subpopulation.
- 39 -
Figure 17. The logistic response to precipitation during the driest quarter by the Louisiana subpopulation.
- 40 -
Figure 18. The logistic response to precipitation during the wettest month by the Louisiana subpopulation.
- 41 -
Figure 19. The logistic response to the mean temperature during the driest quarter by the Louisiana subpopulation.
- 42 -
Figure 20. Test gain for the model run with Texas-Louisiana as a combined species.
- 43 -
Figure 21. Average probability distribution map from the Maxent model run with Texas- Louisiana subpopulations as a combined species.
- 44 -
Figure 22. Test gain for the model run with Texas as an individual species.
- 45 -
Figure 23. Average probability distribution map from the Maxent model run with Texas subpopulation as an individual species.
- 46 -
Figure 24. Test gain for the model run with Louisiana as an individual species.
- 47 -
Figure 25. Average probability distribution map from the Maxent model run with Louisiana as an individual species.
- 48 -
Chapter 4: Discussion
2014 East Texas Survey Discussion
The 2014 survey of the historic sites in the Cypress Creek drainage found 14 of the 25 sites to possibly be missing O. maletae. The organism may still be present in these sites, however, the population numbers may have reached such low numbers as to be undetectable using traditional trapping and netting methods. The extent of the trapping and dip-netting lends confidence to the fact that this organism is either in decline or absent all together from these 14 sites. Individuals were, however, detected in 10 of the historic sites.
We can infer from the presence and absence data that this species is in decline in East
Texas. Heavy channelization, pollution, and the presence of invasive species, Procambarus clarkii
(Barbaresi 2004), were noted in many of the sites. These factors could be at least partially responsible for the absence or decline of the subpopulations in these sites (Mathews 2008). To determine if there is any correlation between these preliminary observations and decline in the subpopulations in these sites, additional data should be collected to confirm the level of change that has occurred within the habitat range since the original surveys were conducted.
Molecular Ecology Discussion
The molecular data from this analysis supported significant divergence between the Texas and Louisiana subpopulations. The values obtained from the percent sequence divergence for both mitochondrial genes showed divergence to be occurring both between and within the subpopulations. Sequence divergences vary for the COI gene in crayfish with values ranging between 2.42% at the intraspecific level to 15.53% at the interspecific level (Munasinghe 2003).
- 49 -
The percent sequence divergence value obtained in this study was 5.91% between the subpopulations. The 16S gene data also supported divergence between the subpopulations. In crayfish, the 16S gene percent sequence divergence values range from 0.84% at the intraspecific level to 7.63% at the interspecific level (Munasinghe 2003). The percent sequence divergence value obtained in this study was 2.37% between the subpopulations. Therefore, the data from the current study indicate that divergence is occurring on both mitochondrial gene segments between the Texas and Louisiana subpopulations at a level somewhere between the amount typically seen in crayfish for intra- vs. interspecific values.
In the state of Texas, a high level of diversity was observed within the Texas subpopulations. Haplotype A is the most ancestral haplotype, and was found to occur in every county within the Texas population range. Haplotypes B-Haplotype G are more derived haplotypes that were not observed in every collection site. Further data, including more samples from each site, should be collected to get a better idea of the divergence occurring within the Texas subpopulation. Haplotype H was only found to exist in the Louisiana collection site. The Louisiana subpopulation’s haplotype had its own separate network. This means that the haplotypes are not similar enough between the states to be connected with 95% or higher certainty (Clemet 2000).
This supports the trend observed in the percent sequence divergence data that these subpopulations between states of O. maleate have undergone substantial divergence.
The substantial divergence that was observed in both mitochondrial genes is likely the result of the ~150 km separation between the Texas and Louisiana subpopulations. The geographic isolation of O. maletae in each subpopulation has resulted in reproductive isolation. There was little to no gene flow observed between the subpopulations. As a result, unique haplotypes were discovered in each subpopulation.
- 50 -
The pairwise sequence divergence obtained from the AMOVA of both mitochondrial gene segments showed complete gene flow to be occurring within the Texas subpopulation, but almost a total lack of gene flow between the Texas and Louisiana subpopulations. These data indicate that the subpopulations are genetically isolated, and therefore unable to interbreed. This lack of gene flow between the subpopulations likely contributed to the genetic divergence observed in this analysis.
Data obtained from the nuclear gene segment, GADPH, was not significant for any of the
AMOVA analyses obtained in this study. This is not uncommon, as the rate of mutation is much lower in nuclear DNA and thus the gene is much more highly conserved between generations
(Munasinghe 2003).
As a result of low sample size and genetic diversity in the Louisiana subpopulation, the only conclusion that can be drawn from these data is that the subpopulations are dissimilar enough to necessitate separate haplotype networks. A pattern emerges in the sequence data that shows higher numbers of unique haplotypes being discovered when more samples are collected from any one site. The limitations of this study were such that only a small number of samples were able to be collected from one site in Louisiana. This led to only one haplotype being discovered within the representatives in this subpopulation. Further studies should be conducted to obtain a more complete analysis of the haplotype distribution of this organism in comparison to the population in Texas.
Ecological Niche Modeling Discussion
The ecological niche modeling results lend insight into the optimal habitat and possible range of O. maletae, as well as what environmental variables are most important to this organism’s
- 51 -
niche. The individual Maxent model for the subpopulation in East Texas showed areas of suitable habitat in the Cypress Creek as well as the Sabine and Sulfur rivers (Figure 14). The organism is not known to occur in the latter two systems, although they contain potentially suitable habitat.
The AUC for the Maxent model with the subpopulations run as a single species was much lower than the models run as separate species. When the models were run as separate species, the
AUC was shown to be higher in both models. The AUC is a measure of model accuracy, and this information supports the idea that the niches are different enough to merit separate models for a higher model fitness.
The subpopulations are separated by a distance of ~150 km. The habitat suitability map did not indicate any areas of shared niche overlap between the two subpopulations. These data indicate that each subpopulation would likely not be supported in any areas of habitat outside of their known individual ranges. Each subpopulation of O. maletae would likely have a low chance of survival if experimental transplantation took place. Future studies in this area, in which an translocation experiment is conducted, could provide further insight into the habitat suability for both populations of O. maletae but may not be optimal due to population declines.
The subpopulations do not share the same conditions for optimal habitat. As such, the
Maxent model analysis supports the conclusion that the subpopulations of O. maletae are ecologically different. The Maxent model predicted the subpopulation of O. maletae in Texas to be found in drier climate conditions that the subpopulation in Louisiana. In East Texas, the subpopulation of O. maletae was predicted to be found in medium-sized streams. However, in
Louisiana, the subpopulation of O. maletae was predicted to be found in smaller order streams.
Historically, O. maletae has been observed in rivers with clear, flowing water and sandy bottoms.
- 52 -
The observation of this species in muddy water with less oxygen was presumed to be a transient situation and not the optimal habitat for this species (Walls 2009). However, the Maxent models for both subpopulations run individually showed that the optimal habitat conditions for O. maletae is, in fact, sand in Texas but clay or mud in Louisiana. Soil conditions are from the areas surrounding the rivers and streams. As such, they may or may not actually reflect the conditions of the river bottoms themselves. A follow up study should be conducted to establish correlation between the soil types of the river bottoms and the optimal habitat for O. maletae in both Texas and Louisiana.
These results are noteworthy as future collection efforts of this organism occur in both states. The collection methodology may need to be revaluated for each state based on the appropriate habitat conditions. For example, since the Texas subpopulation can be found in wider streams, with deeper bottoms, trapping in the center of the stream may yield better results for collecting than dip netting or electrofishing. Whereas, in Louisiana, dip netting and electrofishing may be more appropriate as the organism is most likely found in shallower streams.
- 53 -
Chapter 5: Conclusions
Based on the combined molecular and ecological niche modeling data, the current study has provided compelling evidence the East Texas and Louisiana subpopulations of O. maletae are unique Evolutionary Significant Units (ESUs) (see below)The molecular data suggest that significant differentiation is occurring based on the mitochondrial AMOVA results, the divergence estimates, and the haplotype network. Both more samples and a greater diversity of molecular markers should be added to this analysis to confirm that this species has diverged into two species that are morphologically identical. This would mean that the East Texas species, which was named for its counterpart in the Kisatchie Bayou, is actually a cryptic species indicating that even though, superficially, organisms from the two subpopulations look morphologically indistinguishable, they are not the same species.
The ecological niche modeling data also supports the idea that allopatric speciation is occurring. The subpopulations of O. maletae do not have the same optimal habitat ranges. They have almost completely different habitat preferences according to the predictive Maxent models.
As such, they share no habitat in common that would be suitable for both populations as predicted by ArcGIS habitat suitability. More evidence should be collected to confirm the predictions of the habitat suitability models, and to find new locations in which the organism may be found. A follow up study in which the accuracy of the models are tested by ground truthing would be appropriate.
Both subpopulations are in decline and have been reported as absent or missing in at least half of their historical sites. Both states have listed this organism as a species of conservation concern, pending the results of the studies occurring in East Texas. The data obtained in this study lends compelling evidence that a recommendation for federal conservation status may be
- 54 -
appropriate for this organism. Only one river system in Texas, Cypress Creek, and one in
Louisiana, the Kisatchie Bayou, are known to support O. maletae. Conservation efforts would already be necessary for this species in each state. However, this study has uncovered evidence of a potential cryptic species of crayfish with an urgent need for conservation efforts in East Texas.
This would also raise the need for conservation efforts in Louisiana to a critical level, as this state only has one river system with suitable habitat and already declining populations of this organism.
An ESU is a diversity measurement that denotes a population is distinct for conservation purposes (Moritz 1994). This analysis has shown the subpopulations of O. maletae to meet criteria for separate ESUs as follows: the subpopulations have: 1) restricted gene flow leading to genetic differentiation, and 2) geographic isolation. As such, O. maletae is a good candidate species for being considered an ESU. However, future studies in which reciprocal transplantation experiments take place might not be appropriate as both subpopulations are already in decline and the organism is rare.
- 55 -
Literature Cited
Adams, S., Schuster, G.A. and Taylor, C.A. Orconectes - 56 -oloniz. The IUCN Red List of
Threatened Species. Version 2014.2. (2010)
Astrin, Jonas J., et al. “Molecular taxonomy in pholcid spiders (Pholcidae, Araneae): evaluation
of species identification methods using CO1 and 16S rRNA.” Zoologica Scripta 35.5
(2006): 441-457.
Barbaresi, Silvia, Elena Tricarico, and Francesca Gherardi. "Factors inducing the intense
burrowing activity of the red-swamp crayfish, Procambarus clarkii, an invasive
species." Naturwissenschaften 91.7 (2004): 342-345.
Bird, Christopher E., et al. “Detecting and measuring genetic differentiation.” Phylogeography
and population genetics in Crustacea 19 (2011): 31-55.
Berrill, Michael, and Michael Arsenault. "Spring breeding of a northern temperate crayfish,
Orconectes rusticus." Canadian Journal of Zoology 60.11 (1982): 2641-2645.
Bromberg, C., et al. "Sequencher." Gene Codes Corporation (1995).
Buhay, Jennifer E., and Keith A. Crandall. “Subterranean phylogeography of freshwater
crayfishes shows extensive gene flow and surprisingly large population sizes.” Molecular
Ecology 14.14 (2005): 4259-4273.
Buhay, Jennifer E., G Moni, N Mann, and KA Crandall. “Molecular taxonomy in the dark:
Evolutionary history, phylogeography, and diversity of cave crayfish in the subgenus
- 56 -
Aviticambarus, genus Cambarus.” Molecular Phylogenetics and Evolution 42.2 (2007):
435-448.
Carter, Lane, Young, and Nicholas. “Identifying Niche Overlap from Maxent Model
Predictions.” Natural Resource Ecology Laboratory. Colorado State University. (2011).
Clement, Mark, D. C. K. A. Posada, and Keith A. Crandall. "TCS: a computer program to
estimate gene genealogies." Molecular ecology 9.10 (2000): 1657-1659.
Clegg, M. T., G. M. Brown Jr, W. Y. Brown, W. L. Fink, J. Harte, O. A. Houck, and M. Lynch.
Science and the Endangered Species Act. Washington DC: National Research Council
(1995).
Crandall, Keith A., Henry W. Robison, and Jennifer E. Buhay. “Avoidance of extinction through
nonexistence: the use of museum specimens and molecular genetics to determine the
taxonomic status of an endangered freshwater crayfish.” Conservation Genetics 10.1
(2009): 177-189.
Elith, Jane, Mark A. Burgman, and Helen M. Regan. "Mapping epistemic uncertainties and
vague concepts in predictions of species distribution." Ecological modelling 157.2
(2002): 313-329.
Esri. "ArcGIS 10.3 for Desktop." (2014).
Excoffier, Laurent, Peter E. Smouse, and Joseph M. Quattro. “Analysis of molecular variance
inferred from metric distances among DNA haplotypes: application to human
mitochondrial DNA restriction data.” Genetics 131.2 (1992): 479-491.
- 57 -
Faber-Langendoen, D., L. Master, J. Nichols, K. Snow, A. Tomaino, R. Bittman, G. Hammerson,
B. Heidel, L. Ramsay, and B. Young. “NatureServe conservation status assessments:
methodology for assigning ranks.” NatureServe, Arlington, Virginia (2009).
Fetzner, James W., and Keith A. Crandall. “Linear habitats and the nested clade analysis: an
empirical evaluation of geographic versus river distances using an Ozark crayfish
(Decapoda: Cambaridae).” Evolution 57.9 (2003): 2101-2118.
Fitzpatrick, J. F., “The subgenera of the crawfish genus Orconectes (Decapoda: Cambaridae).”
Proceedings of the Biological Society of Washington 100.1 (1987): 44-74
Garvey, James E., Roy A. Stein, and Heather M. Thomas. “Assessing how fish predation and
interspecific prey competition influence a crayfish assemblage.” Ecology (1994): 532-
547.
Ghia, Daniela, et al. “Modelling environmental niche for the endangered crayfish
Austropotamobius pallipes complex in northern and central Italy.” Freshw Crayfish 19
(2013): 189-195.
Hebert, Paul DN, Alina Cywinska, and Shelley L. Ball. "Biological identifications through DNA
barcodes." Proceedings of the Royal Society of London B: Biological Sciences 270.1512
(2003): 313-321.
Henikoff, S., Henikoff, J.G., “Position-based sequence weights.” Journal of Molecular Biology.
1994 Nov 4;243(4):574-8.
Hobbs, H. H., Lodge, D. M. eds. “Ecology and Classification of North American freshwater
invertebrates.” Chapter 22 (2010): 901-967.
- 58 -
Lake, P. S., Palmer, Margaret A., Biro, Peter, Cole, Jonathan, Covich, Alan P., Dahm, Cliff,
Gibert, Janine, Goedkoop, Willem, Martens, Koen, and Verhoeven, Jos. “Global Change
and the Biodiversity of Freshwater Ecosystems: Impacts on Linkages between Above-
Sediment and Sediment Biota All forms of anthropogenic disturbance—changes in land
use, biogeochemical processes, or biotic addition or loss—not only damage the biota of
freshwater sediments but also disrupt the linkages between above-sediment and sediment-
dwelling biota.” BioScience 50.12 (2000): 1099-1107.
Lodge, David M.,Kershner, Mark W., Aloi, Jane E., and Covich, Alan P. “Effects of an
omnivorous crayfish (Orconectes rusticus) on a freshwater littoral food web.” Ecology
(1994): 1265-1281.
Maddison, Wayne P., and David R. Maddison. "Mesquite: a modular system for evolutionary
analysis." (2001).
Mathews, L. M., Schubart, C. D., Neigel J. E., and Felder, D. L. “Genetic, ecological, and
behavioural divergence between two sibling snapping shrimp species (Crustacea:
Decapoda: Alpheus).” Molecular Ecology 11.8 (2002): 1427-1437.
Mathews, Lauren M., Adams, Luke, Anderson, Erica, Basile, Matthew, Gottardi, Elizabeth, and
Buckholt, Michael A. “Genetic and morphological evidence for substantial hidden
biodiversity in a freshwater crayfish species complex.” Molecular Phylogenetics and
Evolution 48.1 (2008): 126-135.
Mariani, Stefano, et al. “North Sea herring population structure revealed by microsatellite
analysis.” Marine Ecology Progress Series (2005).
- 59 -
Moritz, Craig. "Defining ‘evolutionarily significant units’ for conservation." Trends in ecology &
evolution 9.10 (1994): 373-375.
Munasinghe, DH N., N. P. Murphy, and C. M. Austin. “Utility of mitochondrial DNA sequences
from four gene regions for systematic studies of Australian freshwater crayfish of the
genus Cherax (Decapoda: Parastacidae).” Journal of Crustacean Biology 23.2 (2003):
402-417.
National Research Council (US). Committee on Scientific Issues in the Endangered Species Act.
Science and the endangered species act. National Academies Press, 1995.
Phillips, S. J. 2005. A brief tutorial on Maxent. AT&T Research.
http://www.cs.princeton.edu/~schapire/maxent/tutorial/tutorial.doc (accessed 3/5/2017).
Phillips, Steven J., Robert P. Anderson, and Robert E. Schapire. "Maximum entropy modeling of
species geographic distributions." Ecological Modelling 190.3 (2006): 231-259.
Phillips, Steven J., and Miroslav Dudík. "Modeling of species distributions with Maxent: new
extensions and a comprehensive evaluation." Ecography 31.2 (2008): 161-175.
Rabeni, Charles F., Collierb, Kevin J., Parkynb, Stephanie M., Hicksc, Brendan J.. “Evaluating
techniques for sampling stream crayfish (Paranephrops planifrons).” New Zealand
Journal of Marine and Freshwater Research 31.5 (1997): 693-700.
Raxworthy, Christopher J., et al. “Applications of ecological niche modeling for species
delimitation: a review and empirical evaluation using day geckos (Phelsuma) from
Madagascar.” Systematic Biology 56.6 (2007): 907-923.
- 60 -
Richman, Nadia I., et al. "Multiple drivers of decline in the global status of freshwater crayfish
(Decapoda: Astacidea)." Philosophical Transactions Royal Society B 370.1662 (2015):
20140060.
Schoener, Thomas W. "The Anolis lizards of Bimini: resource partitioning in a complex fauna."
Ecology 49.4 (1968): 704-726.
Scholtz, G. “Phylogeny and Evolution.” Biology of Freshwater Crayfish. Blackwell Science,
Oxford (2002): 30-52.
Schubart, C. D., J. E. Neigel, and D. L. Felder. “Use of the mitochondrial 16S rRNA gene for
phylogenetic and population studies of Crustacea.” Crustacean Issues 12 (2000): 817-
830.
Sinclair, Elizabeth A., et al. “Proposal to complete a phylogenetic taxonomy and systematic
revision for freshwater crayfish (Astacidea).” Freshwater Crayfish 14 (2004): 21-29.
Taylor, Christopher A., and Jason H. Knouft. “Historical influences on genital morphology
among sympatric species: gonopod evolution and reproductive isolation in the crayfish
genus Orconectes (Cambaridae).” Biological Journal of the Linnean Society 89.1 (2006):
1-12.
Taylor, C.A., Schuster, G.A., Cooper, J.E., DiStefano, R.J., Eversole, A.G., Hobbs III, H.H.,
Robison, H.W., Skelton, C.W. and Thoma, R.F. 2007. “A reassessment of the
conservation status of crayfishes of the United States and Canada after 10+ years of
increased awareness.” Fisheries, American Fisheries Society 32(8): 372-389.
- 61 -
Thompson, Julie D., Toby Gibson, and Des G. Higgins. "Multiple sequence alignment using
ClustalW and ClustalX." Current protocols in bioinformatics (2002): 2-3.
Usio, Nisikawa. “Effects of crayfish on leaf processing and invertebrate - 62 -olonization of
leaves in a headwater stream: decoupling of a trophic cascade.” Oecologia 124.4 (2000):
608-614.
Usio, Nisikawa, and Colin R. Townsend. “Roles of crayfish: consequences of predation and
bioturbation for stream invertebrates.” Ecology 85.3 (2004): 807-822.
Walls, Jerry G., Crawfishes from the Kisatchie Ranger District, Kisatchie National Forrest,
Natchitoches Parish, Louisiana. Presented in fulfillment of a contract with Mr. David
Byrd, Fisheries Program Manager, KNF, Pineville, LA. (2005)
Walls, J.G. 2009. Crawfishes of Louisiana. Louisiana State University Press, Louisiana.
Warren, Dan L., Richard E. Glor, and Michael Turelli. "Environmental niche equivalency versus
conservatism: quantitative approaches to niche evolution." Evolution 62.11 (2008): 2868-
2883.
Weir, Bruce S., and C. Clark Cockerham. “Estimating F-statistics for the analysis of population
structure.” Evolution (1984): 1358-1370.
Weir, Bruce S., and C. C. Cockerham. “Genetic data analysis II: Methods for discrete population
genetic data. Sinauer Assoc.” Inc., Sunderland, MA, USA (1996).
Wetzer, R. "Hierarchical analysis of mtDNA variation and the use of mtDNA for
isopod." Contributions to Zoology 70.1 (2001): 23-39.
- 62 -
Wilcove, David S., D Rothstein, J Dubow, A Phillips, E Losos. “Quantifying threats to imperiled
species in the United States.” BioScience (1998): 607-615.
Wright, Sewall. “Isolation by distance.” Genetics 28.2 (1943): 114.
Wright, Sewall. "The relation of livestock breeding to theories of evolution." Journal of Animal
Science 46.5 (1978): 1192-1200.
- 63 -