Morphological evidence for introgressive hybridization between Feirana quadranus and Feirana taihangnica in Qinling Mountains, China
- Published
- Accepted
- Subject Areas
- Evolutionary Studies, Molecular Biology, Taxonomy, Zoology, Statistics
- Keywords
- morphotype, 16S rRNA, principal component analysis, linear discriminant analysis, morphology, morphometric analysis, phylogenetic tree, contact zone, Amphibia
- Copyright
- © 2016 Song et al.
- Licence
- This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ Preprints) and either DOI or URL of the article must be cited.
- Cite this article
- 2016. Morphological evidence for introgressive hybridization between Feirana quadranus and Feirana taihangnica in Qinling Mountains, China. PeerJ Preprints 4:e2058v2 https://doi.org/10.7287/peerj.preprints.2058v2
Abstract
Background. Feirana quadranus and Feirana taihangnica, two species of frogs inhabiting in waterbodies in the Qinling Mountains, China, are believed to be sister species that diverged 4.6 million years ago. In their sympatric area, morphological variations found between the two species imply that the two species had inter-bred. Additionally, F. taihangnica’s polyandrous breeding behavior, without amplexus, would not hinder the potential hybridization. Methods. To verify the hybridization, 117 specimens of F. quadranus and F. taihangnica were collected from eight sampling sites in their sympatric area, and 110 of the specimens were classified morphologically into VV, vw&wv, and ww, representing the putative parental and suspected hybrid types. Their maternal bloodlines were identified using a phylogenetic tree based on a region of the mitochondrial 16S rRNA gene. In total, 34 morphometric indices were selected to analyze the morphological variation between 16S-types or among morphotypes. A principal component analysis (PCA) and linear discriminant analysis (LDA) were conducted on total or partial indices for females, males, and total specimens, as well as simulated populations with falsified morphotypes. The most important indices for differentiation among morphotypes were revealed with the assistance of heat-maps. Results. In the mitochondrial DNA tree, most of the VV were in the same clade as the reference F. quadranus, labeled as Q, while most of the ww and vw&wv were grouped with the reference F. taihangnica, labeled as T. According to the PCA, there was a clear differentiation between VV and ww, while vw&wv specimens were in the middle area close to ww. According to the LDA, VV, vw&wv, and ww were clustered into three separate groups. An ambiguous differentiation between Q and T was shown both in mtDNA tree and in multivariate analyses. Seven of the specimens with conflicting classifications blurred the morphological boundary between Q and T. In both the PCA and LDA, indices that were based on the extent of bumps and skin coloration discriminated VV, vw&wv, and ww better than ratio indices that were derived from measurements. Discussion. The distribution of VV, vw&wv, and ww in multivariate spaces, especially vw&wv being scattered between VV and ww, demonstrated an introgressive hybridization pattern. The extents of bumps in the shape of an inverted "V" between the shoulder blades, spot pattern on the back, and large bumps above the anal region were the most important characteristics for differentiating between three morphotypes or between F. quadranus and F. taihangnica.
Author Comment
1. "46,000 years" in the sentence " Feirana quadranus and Feirana taihangnica, two species of frogs inhabiting in waterbodies in the Qinling (Tsinling) Mountains, China, are believed to be sister species that diverged 46,000 years ago", which was a mistake, has been changed into 4.6 Million years; 2. All the names of Tsinling Mountains in the main article and on the map have been changed into the more updated name "Qinling Mountains"; 3. A green zone representing the range of Feirana kangxianensis has been added on the map; 4. The format in the reference section has been perfected. For example, latin names of species have been italicized, and some of the abbreviations for journals' names following the full version have been deleted.
Supplemental Information
Estimates of evolutionary divergence over sequence pairs (means ± standard errors of p-distance), between and within groups
a Q represents samples genetically classified as F. quadaranus. b T represents samples genetically classified as F. taihangnica. c reference F. quadaranus, F. kangxianensis, or F. taihangnica
Raw data of morphometric measurements
It is also available at https://figshare.com/s/be10cb912d81f0c419aa
The mean ± standard deviation and the difference in the means (p-values) of 34 morphometric indices for 110 specimens were grouped in five ways
"16S" indicates being grouped by the 16S rRNA gene (Q and T); "morp3" indicates being grouped by the three morphotypes (VV, vw&wv, and ww); "sex" indicates being grouped by the sexes (female and male); "F_morp3" indicates the females being grouped together and "M_morp3" indicates the males being grouped together. P-values were calculated by Bartlett’s test, ANOVA, Kruskar–Wallis test, and the t-test. The significance level for each index in each group is estimated using the p-value of t-test as the first priority; if there is no t-test result, then the ANOVA is used, unless the p-value of Bartlett’s test is below 0.05, as marked in red, meaning that the variance of the means is significant different, and thus, the ANOVA is not applicable. In this situation, the p-value of the Kruskar–Wallis test is used instead. Red highlight and postfixed ***: p-value < 1.0E-09; red highlight and postfixed **: 1.0E-09 ≤ p-value < 0.01; yellow highlight and postfixed *: 0.01 ≤ p-value < 0.1; green highlight and postfixed `: 0.1 ≤ p-value < 0.3; and blue highlight and no postfix: p-value > 0.5. It is also available at https://figshare.com/s/be10cb912d81f0c419aa
The p-values that estimate the significance level in Table S3 for the five groups
"16S" column and "sex" column calculated by t-tests, refer to the difference in means between Q and T, and between sexes, respectively; "morp3" column, "F_morp3" column and "M_morp3" column calculated by ANOVA or Kruskar–Wallis tests, refer to the differences in means among the three morphotypes (VV, vw&wv, and ww) in all 110 specimens, in 52 females, and in 58 males, respectively. Red highlight and red text: p-value < 1.0E-09; red highlight and dark red text: 1.0E-09 ≤ p-value < 0.01; yellow highlight: 0.01 ≤ p-value < 0.1; green highlight: 0.1 ≤ p-value < 0.3; and blue highlight: p-value > 0.3. See plot version in Fig. S1.
Means and standard deviations (mean ± s.d.) of morphometric indices for the three morphotypes in both female and male sets
a sample sizes are shown in parentheses. b Dark, medium, and light pink represent the highest, middle, and lowest values, respectively, for each index. c ***: p-value < 1.0E-09; **: 1.0E-09 ≤ p-value < 0.01; *: 0.01≤ p-value < 0.1; `: 0.1 ≤ p-value < 0.3; and none: p-value > 0.3. The p-values correspond to those in the last two columns of Table S4 (for details see Table S3). d Measured characteristics: SVL: snout-vent length, used to eliminate body size effects; HL: head length, from posterior end of mandible to tip of snout; HW: head width, measured at corners of the mouth; SL: snout length, distance between anterior edge of orbit and tip of snout; NED: nostril-to-eye distance, distance between centre of nostril and anterior edge of orbit; IND: internarial distance, distance between inner ends of nostrils; IOD: interorbital distance, shortest distance between inner edges of upper eyelids; IAE: distance between anterior corners of eyes; IPE: distance between posterior corners of eyes; LHL: length of lower arm and hand, from elbow to tip of third finger; HAL: hand length, from base of outer palmar tubercle to tip of third finger; TEL: femur length, from vent to knee; TL: tibia length; TFL: tibiofibula length (length of tarsus and foot), from base of tarsus to tip of fourth toe; FL: foot length, from proximal end of inner metatarsal tubercle to tip of fourth toe; T5FFL: length of free flap of the fifth toe, length of cutaneous fringe along the outer margin of the fifth toe; F1L: first finger length, from proximal end of thenar tubercle to tip of first finger. F3L: partial 3rd finger length, distance between basal border of third finger to tip of third finger; F4L: fourth finger length, from proximal end of thenar tubercle to tip of fourth finger. e Extent indices (scored between 0, indicating the trait was not seen, and 1, indicating the maximum extent): BBE: extent of big bumps above the anal region; SBE: extent of small bumps above the anal region; VBE: extent of bumps in the shape of inverted "V" between shoulder blades; VSE: extent of patch in the shape of inverted "V" between shoulder blades; LBE: extent of line-shaped bumps on the back; BSE: the extent of spot pattern on the back; LSE: extent of strip or or spot pattern on legs (scored between 0, indicating no obvious pattern, 1, indicating pure strip pattern, and 2, indicating pure spot pattern).
Maximum likelihood phylogenetic tree based on 16S rRNA gene partial sequences
The bootstrap support values are shown below branches (The percentage of trees in which the associated taxa clustered together is shown below the branches). Scale bar indicates an evolutionary distance of 0.05 nucleotides per position in the sequence (The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.). Feirana specimens are named by a combination of voucher number and corresponding morphotype, the F. quadranus, F. taihangnica and F. kangxianensis references are named by a combination of GenBank number and species name, and the two R. rugosa outgroups are named by a combination of voucher number and species name. Pink and blue rectangles indicate four specimens with conflicting morphotypical classifications.
Heat-maps of weighted or non-weighted rotation matrices of the PCA (A–D), and coefficient matrices of the linear discriminant functions of the LDA (E–H)
In the weighted matrix, the corresponding proportion of explained variance for each PC or the corresponding proportion of explained discriminability for each LD is in parentheses. Red and blue palettes distinguish between weighted values (red) and non-weighted values (blue). (A) The first 10 PCs for females, corresponding to Fig. 5A; (B) The first 10 PCs for males, corresponding to Fig. 5E; (C) The first 10 PCs for the total specimens, corresponding to Fig. 5B; (D) The first two PCs of the extent indices for the total specimens, corresponding to Fig. 5H; (E) LD for females, grouped in three morphotypes, corresponding to Fig. 6A; (F) LD for males, grouped in three morphotypes, corresponding to Fig. 6E; (G) LD for the total specimens, grouped in three morphotypes, corresponding to Fig. 6B; (H) LD for the total specimens, grouped in five morphotypes, corresponding to Fig. 6F; (I) LD of the 27 ratio indices, grouped in three morphotypes, corresponding to Fig. 6D; (J) LD of the 7 extent indices, grouped in three morphotypes, corresponding to Fig. 6H.