**3. Analysis of genetic diversity and population structure**

### **3.1. Study sites**

In Section 3, genetic diversity and population structure of populations of the Hotoke loach in the upper Kokai R. including 4 adjacent rivers, the southeast part of Tochigi Pref. (Fig. 3) was detailed using the microsatellite loci developed in Section 2 (Table 1). As mentioned in Section 1, populations of the Hotoke loach have been often diminished and isolated by land consolidation projects in rural areas. Therefore it appears difficult to find populations dis‐ tributed with a certain area. However, rich biota still continues to exist in the upper Kokai R. due to delay of land consolidation. This area sounds attractive for field scientists, and then their some activities were carried out to conserve and recover such a sound rural ecosystem

[48-52]. According to the results of these studies [48, 49], the populations of the Hotoke loach tended to be distributed in the upper zone of hill-bottom valleys in this area and also a negative correlation was observed between the population size and water temperature.

**Photo 3.** Collection of individuals of the Hotoke loach in an earth ditch at the site K9 in the Kokai River (unpublished

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach

http://dx.doi.org/10.5772/53022

357

Total genomic DNA from the preserved caudal fins of each individual was extracted using an automated DNA isolation system following the manufacturer's instructions, and kept at

The microsatellite DNA analysis were performed using the following 11 loci that are *Lec01*, *Lec05*, *Lec06*, *Lec08*, *Lec12*, *Lec14*, *Lec15*, *Lec16*, *Lec17*, *Lec18* and *Lec19* with gray color in Table 1. These loci were confirmed to be appropriate for investigating the populations in the Kokai R. in the preliminary studies [37, 38]. In accordance with the procedure in Section 2, micro‐ satellite amplification with PCR on iCycler and C1000 of thermal cyclers was conducted in 10 μl reaction volumes containing approximately 10 ng DNA templates. PCR products were electrophoresed on a 3130*xl* Genetic Analyzer with GeneScan 500 LIZ of size markers and the electrophoregrams were analyzed with the GENEMAPPER. Consequently genotype da‐ ta composed of a pair of fragment sizes, which are inherited from both of parents and de‐ pends on length of repeat motif, was obtained for each individual in a PCR product of a locus. All genotype data were compiled in the software THE EXCEL MICROSATELLITE

The genetic diversity within the populations of the Hotoke loach in each collection site was evaluated with the genotype data of the 11 loci for all individuals. The number of allele (*NA*) and allelic richness (*Ar*) [54], where bias caused by population size (the number of individu‐ als) is removed from *NA*, were estimated using the software GENALEX version 6.41 [55] and FSTAT version 2.9.3 [56], respectively. Differences of *NA* and *Ar* among the populations were

photo)

TOOLKIT [53].

**3.4. DNA data analysis**

*3.4.1. Genetic diversity within population*

**3.3. DNA chemical analysis**

4 °C after being diluted to 10 ng/μl.

**Figure 3.** Collection sites for individuals of populations of the Hotoke loach in the upper of Kokai River (K1 to K20) along with adjacent the Oh, Sakura, Gogyo and Ara Rivers (O, S, G and A1, A2, respectively), the southeast part of Tochigi Prefecture (unpublished figure).

Considering such spatial distribution patterns in the previous studies [48, 49] and geograph‐ ical conditions in this area, a total of 20 sites were established to collect individuals of the populations in the upper Kokai R. (K1 to K20 in Fig. 3). Additionally 5 collection sites of ad‐ jacent 4 rivers that are the Oh, Sakura, Gogyo and Ara (O, S, G and A1, A2, respectively in Fig. 3) were decided to compare with the populations of the Kokai R.

### **3.2. Sample collection**

Sample collections in each site (Fig. 3) were performed using hand nets with reticulation at 2 mm, flame width at 30 to 40 cm in August 2007 to June 2008 (Photo 4). 10 to 24 individuals (a total of 573 individuals) of each population were collected in earth canals and ditches with water depth of 2 to 24 cm, water width of 15 to 110 cm, flow velocity of 5 to 25 cm/s and substrates consisting of silts, sands and gravels. There were no rain during the sample collections and a part of the caudal fin (3 mm × 3 mm) of each individual was removed and preserved in 99.5% EtOH at the sites, and then all individuals were immediately released alive. The preserved caudal fins were kept at -30 °C and the mean ± standard deviation in body length for all individuals was 46 ± 11 mm.

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach http://dx.doi.org/10.5772/53022 357

**Photo 3.** Collection of individuals of the Hotoke loach in an earth ditch at the site K9 in the Kokai River (unpublished photo)

### **3.3. DNA chemical analysis**

[48-52]. According to the results of these studies [48, 49], the populations of the Hotoke loach tended to be distributed in the upper zone of hill-bottom valleys in this area and also a negative correlation was observed between the population size and water temperature.

An Integrated View of the Molecular Recognition and Toxinology - From Analytical Procedures to Biomedical

**Figure 3.** Collection sites for individuals of populations of the Hotoke loach in the upper of Kokai River (K1 to K20) along with adjacent the Oh, Sakura, Gogyo and Ara Rivers (O, S, G and A1, A2, respectively), the southeast part of

Considering such spatial distribution patterns in the previous studies [48, 49] and geograph‐ ical conditions in this area, a total of 20 sites were established to collect individuals of the populations in the upper Kokai R. (K1 to K20 in Fig. 3). Additionally 5 collection sites of ad‐ jacent 4 rivers that are the Oh, Sakura, Gogyo and Ara (O, S, G and A1, A2, respectively in

Sample collections in each site (Fig. 3) were performed using hand nets with reticulation at 2 mm, flame width at 30 to 40 cm in August 2007 to June 2008 (Photo 4). 10 to 24 individuals (a total of 573 individuals) of each population were collected in earth canals and ditches with water depth of 2 to 24 cm, water width of 15 to 110 cm, flow velocity of 5 to 25 cm/s and substrates consisting of silts, sands and gravels. There were no rain during the sample collections and a part of the caudal fin (3 mm × 3 mm) of each individual was removed and preserved in 99.5% EtOH at the sites, and then all individuals were immediately released alive. The preserved caudal fins were kept at -30 °C and the mean ± standard deviation in

Fig. 3) were decided to compare with the populations of the Kokai R.

body length for all individuals was 46 ± 11 mm.

Tochigi Prefecture (unpublished figure).

Applications

356

**3.2. Sample collection**

Total genomic DNA from the preserved caudal fins of each individual was extracted using an automated DNA isolation system following the manufacturer's instructions, and kept at 4 °C after being diluted to 10 ng/μl.

The microsatellite DNA analysis were performed using the following 11 loci that are *Lec01*, *Lec05*, *Lec06*, *Lec08*, *Lec12*, *Lec14*, *Lec15*, *Lec16*, *Lec17*, *Lec18* and *Lec19* with gray color in Table 1. These loci were confirmed to be appropriate for investigating the populations in the Kokai R. in the preliminary studies [37, 38]. In accordance with the procedure in Section 2, micro‐ satellite amplification with PCR on iCycler and C1000 of thermal cyclers was conducted in 10 μl reaction volumes containing approximately 10 ng DNA templates. PCR products were electrophoresed on a 3130*xl* Genetic Analyzer with GeneScan 500 LIZ of size markers and the electrophoregrams were analyzed with the GENEMAPPER. Consequently genotype da‐ ta composed of a pair of fragment sizes, which are inherited from both of parents and de‐ pends on length of repeat motif, was obtained for each individual in a PCR product of a locus. All genotype data were compiled in the software THE EXCEL MICROSATELLITE TOOLKIT [53].

### **3.4. DNA data analysis**

### *3.4.1. Genetic diversity within population*

The genetic diversity within the populations of the Hotoke loach in each collection site was evaluated with the genotype data of the 11 loci for all individuals. The number of allele (*NA*) and allelic richness (*Ar*) [54], where bias caused by population size (the number of individu‐ als) is removed from *NA*, were estimated using the software GENALEX version 6.41 [55] and FSTAT version 2.9.3 [56], respectively. Differences of *NA* and *Ar* among the populations were

tested by one-way analysis of variance (ANOVA) using the software EKUSERU-TOUKEI 2010 (Social Survey Research Information Co., Ltd.).

runs for each cluster were organized using the software STRUCTURE HARVESTER web version 0.6.92 [76] and then summarized using the software CLUMPP [77]. When individu‐ als had the values of Q more than 0.7, they were assigned to be members of that particular cluster in this study. And also K usually appears to show the genetic structure at the upper‐ most hierarchical level [75]. Therefore, when a particular cluster was formed by some popu‐ lations, additional analysis of each cluster was performed to investigate the detailed genetic

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach

All the 11 microsatellite loci were moderate to highly polymorphic, with the number of al‐ leles (*NA*) and observed and unbiased heterozygosities (*HO* and *HE*, respectively) per locus for all individuals ranging from 2 (*Lec19*) to 40 (*Lec06*) and from 0.147 (*Lec17* and *Lec19*) to 0.846 (*Lec05*) and from 0.155 (*Lec17*) and 0.915 (*Lec08*), respectively. Such a polymorphic level observed in these loci indicated to be beneficial to investigating genetic characteristics of

Means of *NA* per locus in the populations varies from 4.5 (Population A2, hereafter Pop A2) to 8.0 (Pop K11). Allele richness (*Ar*) per locus was standardized by the minimum size of the pop‐ ulation (10 individuals of Pop G) and its means per locus varied from 3.7 (Pop A2) to 5.6 (Pop K11) among populations (Fig. 4). The one-way analysis of variance (ANOVA) showed that sig‐

**Figure 4.** Means and standard errors of the number of alleles (*NA*) and allelic richness (*Ar*) per locus in the populations (unpublished figure). *Ar* was standardized by the minimum size of the population (10 individuals of Population G) and

there were no significant differences among the populations for both *NA* and *AR* (*p* > 0.05).

*<sup>250</sup>* = 0.641, *MSE* = 11.937, *p* > 0.05 for *NA* and *F24, 250* = 0.459, *MSE* = 4.803, *p* > 0.05 for *Ar*).

were not confirmed among populations (*F24,*

http://dx.doi.org/10.5772/53022

359

structures after the first analysis.

*3.5.1. Genetic diversity within populations*

nificant differences of the means of both *NA* and *Ar*

**3.5. Results and discussions**

populations in detail.

The observed and unbiased expected heterozygosities (*HO* and *HE*, respectively) [57] were calculated by GENALEX [55]. The software ARLEQUIN version 3.11 [58] was used to test deviations from Hardy-Weinberg equilibrium with Fisher's exact probability test, which was run through 100,000 iterations using the Markov chain Monte Carlo (MCMC). Signifi‐ cance values (α = 0.05) of a multiple test were corrected following the sequential Bonferroni procedure [47]. Significant differences of *HO* and *HE* among the populations were detected by one-way ANOVA using EKUSERU-TOUKEI 2010.

### *3.4.2. Genetic population structure among populations*

Genetic population structure among the populations in the Kokai R. including 4 adjacent rivers was elucidated with three analytical methods based on the assumption that mutation of alleles in each locus confirmed to an infinite allele model [59, 60].

First, genetic differentiation between the populations was evaluated with classical pairwise *FST* statistics [61] using ARLEQUIN [58]. Statistical significance (α = 0.05) for values of *FST* was test‐ ed with applying 10,000 permutations, followed by sequential Bonferroni corrections [47] and these values were graded on four classifications for genetic differentiation in the previous study [62]. An analysis of molecular variance (AMOVA) [63] for *FST* was performed to estimate hierarchical genetic structure across the populations. In this AMOVA, the populations were divided into 2 to 6 groups according to geographical condition such as rivers and the distances among collection sites. And then variances among groups, among populations within groups, among individuals within populations and within all individuals were computed for 3 cases of genetic structure using GENALEX [55] with 10,000 permutations.

Second, a phylogenetic tree of a genetic distance *DA* [64] between the populations was con‐ structed with the neighbor-joining method [65] and the reliability of the obtained phyloge‐ netic tree was evaluated using the aid of 1,000 bootstrap replicates [66]. The software POPULATIONS version 1.2.31 [67] was used to estimate *DA* and to construct a phylogenetic tree and an appropriate shape of the phylogenetic tree was edited with the software MEGA version 5.05 [68].

Finally, Bayesian cluster analysis [69-73] that has been recently used as a popular method was implemented in the software STRUCTURE version 2.3.3 [70] to circumstantially investi‐ gate the occurrence of genetic structure among the populations without the prior identifica‐ tion of populations. Briefly, this analysis allows the inference of the number of genetically homogeneous clusters (K) that are implicitly genetic populations from individual genotypes at multiple loci and also assignment probability (Q) of individuals to each genetic cluster. The admixture model and correlated allele frequencies model were used along with LOCP‐ RIOR model [74] and the software was run with 20 repetitions of 500,000 iterations of MCMC, following a burn-in of 500,000 iterations at K of 1 to 10.

The most likely number of genetic clusters was evaluated using the rate of change in the log probability between the values of successive K [75]. Distribution of the values of Q across runs for each cluster were organized using the software STRUCTURE HARVESTER web version 0.6.92 [76] and then summarized using the software CLUMPP [77]. When individu‐ als had the values of Q more than 0.7, they were assigned to be members of that particular cluster in this study. And also K usually appears to show the genetic structure at the upper‐ most hierarchical level [75]. Therefore, when a particular cluster was formed by some popu‐ lations, additional analysis of each cluster was performed to investigate the detailed genetic structures after the first analysis.

#### **3.5. Results and discussions**

tested by one-way analysis of variance (ANOVA) using the software EKUSERU-TOUKEI

An Integrated View of the Molecular Recognition and Toxinology - From Analytical Procedures to Biomedical

The observed and unbiased expected heterozygosities (*HO* and *HE*, respectively) [57] were calculated by GENALEX [55]. The software ARLEQUIN version 3.11 [58] was used to test deviations from Hardy-Weinberg equilibrium with Fisher's exact probability test, which was run through 100,000 iterations using the Markov chain Monte Carlo (MCMC). Signifi‐ cance values (α = 0.05) of a multiple test were corrected following the sequential Bonferroni procedure [47]. Significant differences of *HO* and *HE* among the populations were detected

Genetic population structure among the populations in the Kokai R. including 4 adjacent rivers was elucidated with three analytical methods based on the assumption that mutation

First, genetic differentiation between the populations was evaluated with classical pairwise *FST* statistics [61] using ARLEQUIN [58]. Statistical significance (α = 0.05) for values of *FST* was test‐ ed with applying 10,000 permutations, followed by sequential Bonferroni corrections [47] and these values were graded on four classifications for genetic differentiation in the previous study [62]. An analysis of molecular variance (AMOVA) [63] for *FST* was performed to estimate hierarchical genetic structure across the populations. In this AMOVA, the populations were divided into 2 to 6 groups according to geographical condition such as rivers and the distances among collection sites. And then variances among groups, among populations within groups, among individuals within populations and within all individuals were computed for 3 cases of

Second, a phylogenetic tree of a genetic distance *DA* [64] between the populations was con‐ structed with the neighbor-joining method [65] and the reliability of the obtained phyloge‐ netic tree was evaluated using the aid of 1,000 bootstrap replicates [66]. The software POPULATIONS version 1.2.31 [67] was used to estimate *DA* and to construct a phylogenetic tree and an appropriate shape of the phylogenetic tree was edited with the software MEGA

Finally, Bayesian cluster analysis [69-73] that has been recently used as a popular method was implemented in the software STRUCTURE version 2.3.3 [70] to circumstantially investi‐ gate the occurrence of genetic structure among the populations without the prior identifica‐ tion of populations. Briefly, this analysis allows the inference of the number of genetically homogeneous clusters (K) that are implicitly genetic populations from individual genotypes at multiple loci and also assignment probability (Q) of individuals to each genetic cluster. The admixture model and correlated allele frequencies model were used along with LOCP‐ RIOR model [74] and the software was run with 20 repetitions of 500,000 iterations of

The most likely number of genetic clusters was evaluated using the rate of change in the log probability between the values of successive K [75]. Distribution of the values of Q across

2010 (Social Survey Research Information Co., Ltd.).

Applications

358

by one-way ANOVA using EKUSERU-TOUKEI 2010.

of alleles in each locus confirmed to an infinite allele model [59, 60].

genetic structure using GENALEX [55] with 10,000 permutations.

MCMC, following a burn-in of 500,000 iterations at K of 1 to 10.

version 5.05 [68].

*3.4.2. Genetic population structure among populations*

### *3.5.1. Genetic diversity within populations*

All the 11 microsatellite loci were moderate to highly polymorphic, with the number of al‐ leles (*NA*) and observed and unbiased heterozygosities (*HO* and *HE*, respectively) per locus for all individuals ranging from 2 (*Lec19*) to 40 (*Lec06*) and from 0.147 (*Lec17* and *Lec19*) to 0.846 (*Lec05*) and from 0.155 (*Lec17*) and 0.915 (*Lec08*), respectively. Such a polymorphic level observed in these loci indicated to be beneficial to investigating genetic characteristics of populations in detail.

Means of *NA* per locus in the populations varies from 4.5 (Population A2, hereafter Pop A2) to 8.0 (Pop K11). Allele richness (*Ar*) per locus was standardized by the minimum size of the pop‐ ulation (10 individuals of Pop G) and its means per locus varied from 3.7 (Pop A2) to 5.6 (Pop K11) among populations (Fig. 4). The one-way analysis of variance (ANOVA) showed that sig‐ nificant differences of the means of both *NA* and *Ar* were not confirmed among populations (*F24, <sup>250</sup>* = 0.641, *MSE* = 11.937, *p* > 0.05 for *NA* and *F24, 250* = 0.459, *MSE* = 4.803, *p* > 0.05 for *Ar*).

**Figure 4.** Means and standard errors of the number of alleles (*NA*) and allelic richness (*Ar*) per locus in the populations (unpublished figure). *Ar* was standardized by the minimum size of the population (10 individuals of Population G) and there were no significant differences among the populations for both *NA* and *AR* (*p* > 0.05).

Means of the observed and unbiased expected heterozygosities (*HO* and *HE*, respectively) per locus across all population ranged from 0.418 (Pop A2) to 0.669 (Pop K11) and from 0.507 (Pop A2) to 0.674 (Pop K11), respectively (Fig. 5). Significant departures from the Hardy-Weinberg equilibrium (HWE) were not observed in all the populations. This result indicated that the populations could be applied to the following analyses of genetic population struc‐ ture, because most analyses are often performed under the assumption that population con‐ forms to HWE. The results of one-way ANOVA showed that there were no significant of the differences among the populations for both *HO* and *HE* (*F24, 250* = 0.377, *MSE* = 0.090, *p* > 0.05 for *HO* and *F24, 250* = 0.207, *MSE* = 0.079, *p* > 0.05 for *HE*).

in relatively moderate level. Hence, a serious concern for genetic diversity could not occur in the populations at present. However, there are no confident that such a level of genetic diversity would be sustaining in the future. Monitoring genetic diversity may need including ordinary bi‐

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach

http://dx.doi.org/10.5772/53022

361

The lowest and highest values of *FST* were observed between Pops K15 & K16 and between Pops K18 & A2 (*FST* = 0.008 and 0.246, respectively, Fig. 6). The permutation test showed that all the *FST* were significantly different from zero (*p* > 0.05), except the lowest *FST* between

**Figure 6.** Values of pairwise *FST* between the populations and their grades of genetic differentiation composed of four classifications (unpublished figure). All the value of *FST* were significantly different from zero (*p* > 0.05), except be‐ tween Populations K15 & K16 (*FST* = 0.008). Four classifications of genetic differentiation derive from the previous

Values of *FST* were graded on four classifications for genetic differentiation based on the previ‐ ous study [62]. These classifications imply no, middle, high and extreme genetic differentiation when *FST* ranges from 0 to 0.05, from 0.05 to 0.15, from 0.15 to 0.25 and over 0.25. Applying this grade, 20.3% of the *FST* (32/190) between the populations within the Kokai R. (Pops K1 to K20)

ological investigation such as an estimation of size and age composition of populations.

*3.5.2. Genetic population structure inferred from FST*

study [62].

Pops K15 & K16 after sequential Bonferroni corrections [47].

**Figure 5.** Means and standard errors of the observed and unbiased expected heterozygosities (*HO* and *HE*, respectively) per locus in the populations (unpublished figure). There were no significant differences among the populations for both *HO* and *HE* (*p* > 0.05).

Genetic diversity of the populations appeared not to degrade. Generally, when population size is small, inbreeding among individuals appears to progressively occur in a population [30-33] as mentioned in Section 1. It has been observed that such populations had low values of *NA*, *HO* and *HE* [33]. For instance, means of *NA* per locus for the Ethiopian wolf, *Canis simensis*, the Mauritius kestrel, *Falco punctatus* and the Northern hairy-nosed wombat, *Lasiorhinus krefftii* which are des‐ ignated as worldwide endangered species, were only 2.4, 1.4 and 2.1, respectively. Means of *HE* for the Ethiopian wolf, the Mauritius kestrel and the Northern hairy-nosed wombat were also 0.21, 0.10 and 0.32, respectively [78]. But then, values of representatively common freshwater fish species inhabiting agricultural canals and ditches in rural area, Japan such as the Dojo loach, the Field gudgeon, *Gnathopogon elongates elongatus* and the Amur goby (orange type), *Rhinogo‐ bius* sp. OR ranged from 3.3 to 17.7 (both of the Amur goby) for means of *NA* per locus and from 0.463 (the Dojo loach) to 0.905 (the Field gudgeon) for means of *HE* per locus [79-81].

Comparing with these values for the endangered and common species, the means of *NA* and *HE* per locus (4.5 to 8.0 and 0.507 to 0.674, respectively) observed in the populations indicated to be in relatively moderate level. Hence, a serious concern for genetic diversity could not occur in the populations at present. However, there are no confident that such a level of genetic diversity would be sustaining in the future. Monitoring genetic diversity may need including ordinary bi‐ ological investigation such as an estimation of size and age composition of populations.

### *3.5.2. Genetic population structure inferred from FST*

Means of the observed and unbiased expected heterozygosities (*HO* and *HE*, respectively) per locus across all population ranged from 0.418 (Pop A2) to 0.669 (Pop K11) and from 0.507 (Pop A2) to 0.674 (Pop K11), respectively (Fig. 5). Significant departures from the Hardy-Weinberg equilibrium (HWE) were not observed in all the populations. This result indicated that the populations could be applied to the following analyses of genetic population struc‐ ture, because most analyses are often performed under the assumption that population con‐ forms to HWE. The results of one-way ANOVA showed that there were no significant of the differences among the populations for both *HO* and *HE* (*F24, 250* = 0.377, *MSE* = 0.090, *p* > 0.05 for

An Integrated View of the Molecular Recognition and Toxinology - From Analytical Procedures to Biomedical

**Figure 5.** Means and standard errors of the observed and unbiased expected heterozygosities (*HO* and *HE*, respectively) per locus in the populations (unpublished figure). There were no significant differences among the populations for

Genetic diversity of the populations appeared not to degrade. Generally, when population size is small, inbreeding among individuals appears to progressively occur in a population [30-33] as mentioned in Section 1. It has been observed that such populations had low values of *NA*, *HO* and *HE* [33]. For instance, means of *NA* per locus for the Ethiopian wolf, *Canis simensis*, the Mauritius kestrel, *Falco punctatus* and the Northern hairy-nosed wombat, *Lasiorhinus krefftii* which are des‐ ignated as worldwide endangered species, were only 2.4, 1.4 and 2.1, respectively. Means of *HE* for the Ethiopian wolf, the Mauritius kestrel and the Northern hairy-nosed wombat were also 0.21, 0.10 and 0.32, respectively [78]. But then, values of representatively common freshwater fish species inhabiting agricultural canals and ditches in rural area, Japan such as the Dojo loach, the Field gudgeon, *Gnathopogon elongates elongatus* and the Amur goby (orange type), *Rhinogo‐ bius* sp. OR ranged from 3.3 to 17.7 (both of the Amur goby) for means of *NA* per locus and from

0.463 (the Dojo loach) to 0.905 (the Field gudgeon) for means of *HE* per locus [79-81].

Comparing with these values for the endangered and common species, the means of *NA* and *HE* per locus (4.5 to 8.0 and 0.507 to 0.674, respectively) observed in the populations indicated to be

*HO* and *F24, 250* = 0.207, *MSE* = 0.079, *p* > 0.05 for *HE*).

both *HO* and *HE* (*p* > 0.05).

Applications

360

The lowest and highest values of *FST* were observed between Pops K15 & K16 and between Pops K18 & A2 (*FST* = 0.008 and 0.246, respectively, Fig. 6). The permutation test showed that all the *FST* were significantly different from zero (*p* > 0.05), except the lowest *FST* between Pops K15 & K16 after sequential Bonferroni corrections [47].

**Figure 6.** Values of pairwise *FST* between the populations and their grades of genetic differentiation composed of four classifications (unpublished figure). All the value of *FST* were significantly different from zero (*p* > 0.05), except be‐ tween Populations K15 & K16 (*FST* = 0.008). Four classifications of genetic differentiation derive from the previous study [62].

Values of *FST* were graded on four classifications for genetic differentiation based on the previ‐ ous study [62]. These classifications imply no, middle, high and extreme genetic differentiation when *FST* ranges from 0 to 0.05, from 0.05 to 0.15, from 0.15 to 0.25 and over 0.25. Applying this grade, 20.3% of the *FST* (32/190) between the populations within the Kokai R. (Pops K1 to K20)

were classified into no genetic differentiation and a part of such populations tended to be close located each other (Fig. 6). The remaining *FST* within the Kokai R. were classified into middle ge‐ netic differentiation. Between the populations in the Kokai R. and adjacent 4 rivers (Pops O to A2), their *FST* showed middle to high genetic differentiation, although the *FST* between the popu‐ lations in the Kokai and Oh Rs were partially no differentiation.

Genetic differentiation between the populations was significantly inferred from the analysis of *FST* and its relevant AMOVA. Geographical condition such as river and the distances among locations appeared to relate to degree of the genetic differentiation as illustrated in the previous studies [37, 38, 82-84]. However, only a part of genetic population structure could be indicated in this analysis, because the proportions of the genetic variances at the level of among groups were relatively low (2.8 to 7.3 % of among groups in Table 2). Investi‐ gating schematically and visually genetic structure may have to be implemented as further

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach

http://dx.doi.org/10.5772/53022

363

The calculated genetic distance *DA* between the populations ranged from 0.073 (between Pops K15 & K16) to 0.99 (between Pops K6 & A2). In this phylogenetic tree of *DA* using the neighborjoining method [65] (Fig. 7), there was a few of highly significant divergences of population with the bootstrap probabilities over 90 % (e.g. 98 % between Pops K15 & K16, 93 % between Pops K10 & K11); while the probabilities left were less than 50 % on most divergences. But, the topology of the phylogenetic tree displayed that there were 4 distinct groups, GroupTree 1 to 4 despite weak condition of the statistical support. Both GroupTree 1 and 2 consisted of 3 populations of the Go‐ gyo and Ara Rs (Pops G, A1 and A2) and of the Kokai, Oh and Sakura Rs (Pops K8, O and S), re‐ spectively. GroupTree 3 consisted of 7 populations collected in the lower part of the Kokai R. (Pops K1 to K7), while GroupTree 4 were formed by the 12 remaining populations coming from the mid‐

The schematic genetic structure of the populations was showed by constructing the phylo‐ genetic tree (Fig. 7). Including the results of the above *FST* analysis and AMOVA, the exis‐ tence of 2 genetic populations that related to GroupTree 3 and 4 was indicated in the populations within the Kokai R., but these groups were statistically cryptic. It could be ex‐ pected that characterization of admixture of gene flow and migrants among the populations

The Bayesian clustering analysis supported the occurrence of two defined genetic clusters, Clusters A and B in the uppermost hierarchical level (Fig. 8). By accounting for the number of individuals with more than 70 % of assignment probability (Q) to each cluster, 98.6 6% of all individuals (507/514 individuals) in the populations from the Kokai, Oh and Sakura Rs (Pops K1 to S) were assigned to Cluster A. And also, 91.5 % of the remaining individuals (54/59 individuals) in the populations from the Gogyo and Ara Rs (Pops G to A2) were as‐

Further clustering analysis were performed to assign the populations in Clusters A and B to genetic clusters in the second hierarchical level. Applying the same procedure in the first analysis, the appropriate K were 2 in both analyses. According to the values of Q of the indi‐ viduals, they were assigned to one of Clusters I, II or admixture of Cluster I & II in the anal‐

was displayed by detailing structures of such cryptic genetic populations.

*3.5.4. Genetic structure among populations inferred from Bayesian cluster analysis*

ysis of Cluster A and Clusters III or IV in the analyses of Cluster B.

analysis as commented in the previous study [74].

dle and upper part of the Kokai R. (Pops K9 to K20).

signed to Cluster B (Fig. 8).

*3.5.3. Genetic population structure inferred from phylogentic tree*

The analysis of molecular variance (AMOVA) was implemented for the following Cases I to III, among which the number of groups and composition of the populations in groups differed. In Case I, the populations of the Kokai R. (K1 to K20) and 4 adjacent rivers (Pops O to A2) were div‐ ided into GroupCaseI 1 and 2, respectively. In Case II, GroupCaseII 1 was formed by the populations of the Kokai, Oh and Sakura Rs (Pops K1 to S) and GroupcaseII 2, 3 and 4 were formed by 3 remain‐ ing populations of 2 rivers (Pops G, A1 and A2). There were groups GroupCaseIII 1 to 6 composed of the populations of the Kokai (Pops K1 to K20), Oh (Pop O), Sakura (Pop S), Gogyo (Pop G), one Ara (Pop A1) and another Ara (Pop A2) R. in Case III.

Significant genetic differentiations were observed at all hierarchical levels in all cases (*p* < 0.01, Table 2). The largest genetic variance in all variances was found at the level of within individuals in each case (from 82.5 % in Case II to 86.0 % in Case I). The genetic variances at the levels of among groups and among populations within groups accounted for 2.8% in Case I to 7.3 % in Case II and 6.9 % in Case II and III to 7.9 % in Case I, respectively (Table 2).


a *F*A/E, b*F*B/(B+C+D), c *F*(A+B)/E, d*F*C/(C+E), e *F*(A+B+C)/E

**Table 2.** Results of analysis of molecular variance (AMOVA) for three Cases I, II and III (unpublished table). Compositions of the populations in groups for Case I to III are referred in the text. Genetic differentiations were significant at all hierarchical levels for each case (*p* < 0.01).

Genetic differentiation between the populations was significantly inferred from the analysis of *FST* and its relevant AMOVA. Geographical condition such as river and the distances among locations appeared to relate to degree of the genetic differentiation as illustrated in the previous studies [37, 38, 82-84]. However, only a part of genetic population structure could be indicated in this analysis, because the proportions of the genetic variances at the level of among groups were relatively low (2.8 to 7.3 % of among groups in Table 2). Investi‐ gating schematically and visually genetic structure may have to be implemented as further analysis as commented in the previous study [74].

### *3.5.3. Genetic population structure inferred from phylogentic tree*

were classified into no genetic differentiation and a part of such populations tended to be close located each other (Fig. 6). The remaining *FST* within the Kokai R. were classified into middle ge‐ netic differentiation. Between the populations in the Kokai R. and adjacent 4 rivers (Pops O to A2), their *FST* showed middle to high genetic differentiation, although the *FST* between the popu‐

An Integrated View of the Molecular Recognition and Toxinology - From Analytical Procedures to Biomedical

The analysis of molecular variance (AMOVA) was implemented for the following Cases I to III, among which the number of groups and composition of the populations in groups differed. In Case I, the populations of the Kokai R. (K1 to K20) and 4 adjacent rivers (Pops O to A2) were div‐ ided into GroupCaseI 1 and 2, respectively. In Case II, GroupCaseII 1 was formed by the populations of the Kokai, Oh and Sakura Rs (Pops K1 to S) and GroupcaseII 2, 3 and 4 were formed by 3 remain‐ ing populations of 2 rivers (Pops G, A1 and A2). There were groups GroupCaseIII 1 to 6 composed of the populations of the Kokai (Pops K1 to K20), Oh (Pop O), Sakura (Pop S), Gogyo (Pop G), one

Significant genetic differentiations were observed at all hierarchical levels in all cases (*p* < 0.01, Table 2). The largest genetic variance in all variances was found at the level of within individuals in each case (from 82.5 % in Case II to 86.0 % in Case I). The genetic variances at the levels of among groups and among populations within groups accounted for 2.8% in Case I to 7.3 % in Case II and 6.9 % in Case II and III to 7.9 % in Case I, respectively (Table 2).

> **Among pops within groups (B)**

MS 52.5 16.8 3.4 3.2 Var comp 0.102 0.291 0.124 3.184 3.702 % of var 2.76 7.87 3.35 86.01 100.00 *F* 0.028a 0.081b 0.106c 0.038d 0.140e

MS 34.6 15.9 3.4 3.2 Var comp 0.282 0.267 0.124 3.184 3.858 % of var 7.32 6.93 3.22 82.53 100.00 *F* 0.073 0.075 0.142 0.038 0.175

MS 28.5 15.6 3.4 3.2 Var comp 0.183 0.260 0.124 3.184 3.752 % of var 4.88 6.94 3.31 84.86 100.00 *F* 0.049 0.073 0.118 0.038 0.151

**Table 2.** Results of analysis of molecular variance (AMOVA) for three Cases I, II and III (unpublished table). Compositions of the populations in groups for Case I to III are referred in the text. Genetic differentiations were

**Hierarchy**

d.f. 1 23 548 573 1145

d.f. 3 21 548 573 1145

d.f. 5 19 548 573 1145

**Among inds within pops (C)**

**Within inds (D)** **Total (E)**

lations in the Kokai and Oh Rs were partially no differentiation.

Ara (Pop A1) and another Ara (Pop A2) R. in Case III.

**Statistic**

**Among groups (A)**

**Case (no. groups)**

Applications

362

I (2)

II (4)

III (6)

*F*A/E, b*F*B/(B+C+D), c

*F*(A+B)/E, d*F*C/(C+E), e

*F*(A+B+C)/E

significant at all hierarchical levels for each case (*p* < 0.01).

a

The calculated genetic distance *DA* between the populations ranged from 0.073 (between Pops K15 & K16) to 0.99 (between Pops K6 & A2). In this phylogenetic tree of *DA* using the neighborjoining method [65] (Fig. 7), there was a few of highly significant divergences of population with the bootstrap probabilities over 90 % (e.g. 98 % between Pops K15 & K16, 93 % between Pops K10 & K11); while the probabilities left were less than 50 % on most divergences. But, the topology of the phylogenetic tree displayed that there were 4 distinct groups, GroupTree 1 to 4 despite weak condition of the statistical support. Both GroupTree 1 and 2 consisted of 3 populations of the Go‐ gyo and Ara Rs (Pops G, A1 and A2) and of the Kokai, Oh and Sakura Rs (Pops K8, O and S), re‐ spectively. GroupTree 3 consisted of 7 populations collected in the lower part of the Kokai R. (Pops K1 to K7), while GroupTree 4 were formed by the 12 remaining populations coming from the mid‐ dle and upper part of the Kokai R. (Pops K9 to K20).

The schematic genetic structure of the populations was showed by constructing the phylo‐ genetic tree (Fig. 7). Including the results of the above *FST* analysis and AMOVA, the exis‐ tence of 2 genetic populations that related to GroupTree 3 and 4 was indicated in the populations within the Kokai R., but these groups were statistically cryptic. It could be ex‐ pected that characterization of admixture of gene flow and migrants among the populations was displayed by detailing structures of such cryptic genetic populations.

### *3.5.4. Genetic structure among populations inferred from Bayesian cluster analysis*

The Bayesian clustering analysis supported the occurrence of two defined genetic clusters, Clusters A and B in the uppermost hierarchical level (Fig. 8). By accounting for the number of individuals with more than 70 % of assignment probability (Q) to each cluster, 98.6 6% of all individuals (507/514 individuals) in the populations from the Kokai, Oh and Sakura Rs (Pops K1 to S) were assigned to Cluster A. And also, 91.5 % of the remaining individuals (54/59 individuals) in the populations from the Gogyo and Ara Rs (Pops G to A2) were as‐ signed to Cluster B (Fig. 8).

Further clustering analysis were performed to assign the populations in Clusters A and B to genetic clusters in the second hierarchical level. Applying the same procedure in the first analysis, the appropriate K were 2 in both analyses. According to the values of Q of the indi‐ viduals, they were assigned to one of Clusters I, II or admixture of Cluster I & II in the anal‐ ysis of Cluster A and Clusters III or IV in the analyses of Cluster B.

**Figure 8.** Structures of genetic clusters in the populations inferred by the Bayesian analysis (unpublished figure). Clus‐ ters A & B and I to IV imply the genetic populations at uppermost and second hierarchical levels, respectively. Each individual is represented by a horizontal line fragmented by assignment probabilities to the genetic clusters.

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach

http://dx.doi.org/10.5772/53022

365

In the analysis of Cluster A, 77.0 to 98.6 % of individuals of Pops K2 to K6, K8, O and S (a total of 8 populations) in the Kokai, Oh and Sakura Rs were assigned to members of Cluster I (Fig. 8). Considering the geographical locations of the populations as performed in the pre‐ vious studies [85-88], Cluster I mainly indicated to be the genetic population of the lower

**Figure 7.** Phylogenetic tree of *DA* for the populations with neighbour-joining method (unpublished figure)

**Figure 8.** Structures of genetic clusters in the populations inferred by the Bayesian analysis (unpublished figure). Clus‐ ters A & B and I to IV imply the genetic populations at uppermost and second hierarchical levels, respectively. Each individual is represented by a horizontal line fragmented by assignment probabilities to the genetic clusters.

In the analysis of Cluster A, 77.0 to 98.6 % of individuals of Pops K2 to K6, K8, O and S (a total of 8 populations) in the Kokai, Oh and Sakura Rs were assigned to members of Cluster I (Fig. 8). Considering the geographical locations of the populations as performed in the pre‐ vious studies [85-88], Cluster I mainly indicated to be the genetic population of the lower

**Figure 7.** Phylogenetic tree of *DA* for the populations with neighbour-joining method (unpublished figure)

An Integrated View of the Molecular Recognition and Toxinology - From Analytical Procedures to Biomedical

Applications

364

part of the Kokai including the Oh and Sakura Rs (Fig. 9). 77.4 to 99.0 % of individuals of Pops K9, 10, 12, 13, 15 to 18 (a total of 8 populations) in the Kokai occupied members of Cluster II. Cluster II also implied to be the genetic population of the middle and upper parts of the Kokai R. The remaining individuals of Pops K1, K7, K11, K14, K19 and K20 (a total of 6 populations) in the Kokai R. were mainly classified into members of admixtures of Clus‐ ters I & II. In the analysis of Cluster B, almost all individuals (more than 99.1 %) of Pops G and A1 in the Gogyo and Ara Rs. and Pop A2 in the Ara R. were assigned to Clusters III and IV, respectively (Fig. 8). Cluster III and IV reflected the genetic populations of the Gogyo and one of Ara R. and another of the Ara R., respectively (Fig. 9).

**4. Conclusions**

for conserving the populations.

**Acknowledgement**

A series of the exhaustive genetic analysis in this chapter demonstrated that the populations of the Hotoke loach indicated to have moderate genetic diversity and to be supported with 4 genetic populations, of which distributions depended on the populations and the geographi‐ cal locations. These 2 genetic characteristics showed that there could not be serious genetic concerns at present and the populations might be available as valuable biological resources such as bird food. To fulfill the effective utilization of this loach in near future, both biomed‐ ical and nutritional investigations for component contained in the body may also have to be practiced in the next research subjects along with proposing an optimal management plan

Genetic Diversity and Population Structure of the Hotoke Loach, *Lefua echigonia*, a Japanese Endangered Loach

http://dx.doi.org/10.5772/53022

367

Further, the following 2 suggestions based on the results of this analysis should be realized in conduct of the next research. First, to sustain the present genetic features, habitats of the populations have to be maintained with monitoring the population size. As it is repeatedly described in the above, but reduction of the population size often appears to cause degrada‐ tion of the genetic diversity and the lost genetic diversity could never be regained in the populations [30-33]. Avoiding such a decrease in the genetic diversity, habitat conservation might be important for a population management. It was also investigated that this species had relatively strict water temperature resistance compared with other common freshwater fish [48, 49]; hence the control of the water quality, especially water temperature could be

Second, spatial distribution and composition of the genetic populations should be taken ac‐ count in the population management. In this area the genetic populations could be establish‐ ed by only geographical factors such as river and ground conditions (Fig. 9) and related to no human activities. The foregoing genetic populations often appear a kind of genetic heri‐ tages and it is recommended that their distributions do not have to be disturbed artificially [30-33, 35]. If perchance size diminishment of a specific population is observed and there is only individual translocation as a method to recover the population, selections of translated individuals and populations should be advisedly carried out based on the distribution of the genetic populations. Finally there still may be various and many biological resources left in the rural ecosystem in Japan. Genetic analyses performed in this chapter would have to con‐

Our thanks go to Drs Wataru Kakino and Shin-ichi Matsuzawa, Mr. Masumi Matsuzaki and Ms. Zhenli Gao for their aggressive supports in the research field and Mses. Chikusa Suzuki, Ponthip Goto and Kyoko Yamanoi for their assistance in DNA chemical analyses, including valuable discussions with Dr. Hiroshi Aiki. The first author would like to kindly appreciate Dr. Gandhi Rádis Baptista for his invitation to this book and Mses. Adriana Pecar, Ivana Zec and Masa Vidovic and Mr. Dejan Grgur for their thoughtful helps in the publishing process

tribute substantially to exploration and beneficial utilization of these resources.

one of essential factors for conserving habitats of the populations.

**Figure 9.** Spatial distribution and composition of genetic clusters in populations (unpublished figure). Size of circle reflects that of population.

Consequently the four genetic populations (Clusters I to IV) and a mixed genetic population (admixture of Cluster I & II) were confirmed in the populations using this clustering analy‐ sis. Clusters I, II and a pair of Clusters III & IV nearly coincided with a pair of GroupTree 2 & 3, GroupTree 4 and GroupTree 1 in the phylogenetic tree, respectively (Fig. 7). Moreover, the presence of the mixed genetic population, which could not be usually detected in a phyloge‐ netic tree, was founded by the cluster analysis. As discussed in the previous studies [69-71, 73], this admixture of Cluster I & II may be established through gene flow caused by mi‐ grant; thus, events relative to individual movement and breeding could have occurred among some populations in the past.
