Ecological shifts underlie parallels between ontogenetic and evolutionary allometries in parrotfishes
Data files
Sep 06, 2024 version files 3.93 MB
-
angular_slide.csv
83 B
-
cer_slide.csv
104 B
-
dentary_slide.csv
47 B
-
hyo_slide.csv
132 B
-
Labrid82ModTimeTree.tre
508.91 KB
-
merged_labrid_coords_PARROTS.csv
449.33 KB
-
merged_labrid_coords_WRASSES.csv
950.59 KB
-
merged_labrid_coords.csv
1.68 MB
-
merged_scarus_coords.csv
293.36 KB
-
nasal_slide.csv
63 B
-
neuro_slide_2.csv
365 B
-
ontogeny_groups.csv
15.61 KB
-
pjaw_slide.csv
97 B
-
premax_slide.csv
222 B
-
README.md
12.91 KB
-
Scarines_groups.csv
3.85 KB
-
Scarus_LS.csv
1.66 KB
-
skull_slider_2.csv
1.36 KB
-
uro_slide.csv
102 B
-
Wrasses_groups.csv
7.10 KB
Abstract
During ontogeny, animals often undergo significant shape and size changes, coinciding with ecological shifts. This is evident in parrotfishes (Eupercaria: Labridae), which experience notable ecological shifts during development, transitioning from carnivorous diets as larvae and juveniles to herbivorous and omnivorous diets as adults, using robust beaks and skulls for feeding on coral skeletons and other hard substrates. These ontogenetic shifts mirror their evolutionary history, as parrotfishes are known to have evolved from carnivorous wrasse ancestors. Parallel shifts at ontogenetic and phylogenetic levels may have resulted in similar evolutionary and ontogenetic allometric trajectories within parrotfishes. To test this hypothesis, using micro-CT scanning and 3D geometric morphometrics, we analyze the effects of size on the skull shape of the striped parrotfish Scarus iseri and compare its ontogenetic allometry to the evolutionary allometries of 57 parrotfishes and 162 non-parrotfish wrasses. The young S. iseri have skull shapes resembling non-parrotfish wrasses and grow towards typical adult parrotfish forms as they mature. There was a significant relationship between size and skull shapes, and strong evidence for parallel ontogenetic and evolutionary slopes in parrotfishes. Our findings suggest that morphological changes associated with the ecological shift characterizing interspecific parrotfish evolution are conserved in their intraspecific ontogenies.
README: Ecological shifts underlie parallels between ontogenetic and evolutionary allometries in the striped parrotfish
Continued methods
(a) Ontogenetic allometry of *Scarus iseri*
To analyze the shape variation of entire skull and individual bones of S. iseri across development, we applied Principal Component Analysis (PCA) using gm.prcomp function in geomorph package (Adams et al., 2022).
To estimate coefficients of ontogenetic allometry and test the effects of size on skull and bone shapes (superimposed coords), we used Analysis of Variance, fitting a model with one main effect, the factor “Size”, measured as ln-transformed CS (LCS). The ANOVA was performed using the procD.lm function in geomorph (Collyer and Adams, 2018; 2021). To better visualize these relationships, we predicted individual shapes using group-specific regression models, and the resulting shapes for all samples were plotted using a PCA. We plotted PC1 scores against size. Allometric axes were then plotted on the first PC, along with scores relative to the size axis, using the plotAllometry function in geomorph with the method set to "RegScore"
(b) Ontogenetic and evolutionary allometry comparisons
To compare the ontogenetic allometric coefficients of S. iseri and the evolutionary allometry of parrotfishes, we conducted pairwise analyses involving vectors of allometric coefficients. The growth allometries were estimated using OLS, while evolutionary allometries were estimated using PGLS. Following the methodology outlined by Zelditch and Swiderski (2023), we executed pairwise comparisons by extracting fitted values from the OLS and PGLS models. These values were then combined, and a subsequent ANCOVA was performed using the procD.lm function in the geomorph package. This ANCOVA involved regressing the fitted values against size. Finally, pairwise comparisons between vectors were carried out, utilizing the pairwise function in RRPP.
To explore evolutionary allometry, we used a recently published phylogenetic hypothesis involving 410 species of labrid fishes (Larouche et al., 2023). The tree was pruned down to include the 217 taxa from which morphometric data were collected, a process accomplished using the drop.tip function in the ape package version 5.0 (Paradis and Schliep, 2019). To derive coefficients for evolutionary allometry and assess the evolutionary allometry of parrotfishes, we employed Phylogenetic Analysis of Covariance (PGLS). The model encompassed one principal effect: the factor "Size," measured as the natural logarithm-transformed centroid size (ln-transformed CS or LCS). We used a Type II Sums of Squares for statistical tests of model terms (Adams and Collyer, 2018). In line with the analysis of growth allometry, tests of statistical significance were conducted through permutations of the residuals from the reduced model. The statistical significance of the slopes was determined by comparing it to the distribution of random values. The phylogenetic ANCOVA was fitted using the procD.pgls function in geomorph, and pairwise comparisons between ontogenetic and evolutionary allometries were conducted using the pairwise function in RRPP. Beyond assessing the null hypothesis that the angles differ no more than expected by chance, we also examined the hypothesis that they are no more similar than anticipated by chance. Allometric axes were then plotted on the first PC, along with scores relative to the size axis, using the plotAllometry function in geomorph with the method set to "Predline."
Given the species non-independence we reduced the ordinary least squares (OLS) residuals through PCA up to the first 12 principal components. Subsequently, we estimated Pagels λ for both multivariate datasets, utilizing the transformPhylo.ML function in the motmot package version 2.1.3 (Puttick et al., 2019). Finally, we transformed the phylogeny by the λ value using the rescale function in the geiger package version 2.07 (Pennell et al., 2014).
Description of the data and file structure
We have submitted our museum specimen information (Table_S1.docx
), raw data (landmark coordinate files in a 2d array format for all labrids - merged_labrid_coords.csv
, just for parrotfishes - merged_labrid_coords_PARROTS.csv
, just for wrasses species -merged_labrid_coords_WRASSES.csv
, and just for Scarus iseri - merged_scarus_coords.csv
), slide landmarks for entire skull and each individual bone (skull_slider_2.csv
, angular_slide.csv
, cer_slide.csv
, dentary_slide.csv
, hyo_slide.csv
, nasal_slide.csv
, neuro_slide_2.csv
, pjaw_slide.csv
, premax_slide.csv
, uro_slide.csv
), raw data for group information (ontogeny_groups.csv
, Scarines_groups.csv,
parrots_groups.csv
, wrasses_groups.csv
and Scarus_LS.csv
), raw trimmed phylogenetic trees (Labrid82ModTimeTree.tre
), and R script (Parrotfish_ontogeny.R
).
- Table S1: Species list used in our study (word format): species, sample number and origins (Museum or Fieldwork)
- Species: Currently valid scientific name
- n: sample number
- Source/Museum: Fieldwork Belize, museums (BMNH, ANSP, FMNH, KAUM, UF, UWFC, AM, USNM, AMNH, JFBM, UVI), or Uncat (species not deposited in a museum)
- Catalog number: voucher number associated to the museum or Uncat (species not deposited in a museum)
- Ontogeny - S. iseri shape dataset (merged_scarus_coords.csv): the superimposed landmarks coordinates of 54 S. iseri individuals (csv file) used for ontogenetic analysis (Step 1 of the R script)
- Species (categorical variable): Scarus iseri specimens
- V1-V600 (numerical variable): landmark coordinates in a 2d array formart
- Evolutionary allometry - adult parrotfish shape dataset (merged_labrid_coords_PARROTS.csv): the superimposed landmarks coordinates of 91 parrotfishes (csv file) used for comparing ontogenetic allometry of S. iseri to the evolutionary allometries of Scarinini (Step 2 of the R script)
- Species (categorical variable): Currently valid scientific name
- V1-V600 (numerical variable): landmark coordinates in a 2d array formart
All Labridae evolutionary allometry - adult labrids shape dataset (merged_labrid_coords.csv): the superimposed landmarks coordinates of 315 individuals (csv file) used for comparing ontogenetic allometry of S. iseri to the other Labrids (Step 3 of the R script)
- Species (categorical variable): Currently valid scientific name
- V1-V600 (numerical variable): landmark coordinates in a 2d array format
Wrasses evolutionary allometry - adult labrids shape dataset (merged_labridcoords_WRASSES.csv): the superimposed landmarks coordinates of 176 individuals (csv file) used for comparing ontogenetic allometry of S. iseri to the other Labrids (Step 3 of the R script)
- Species (categorical variable): Currently valid scientific name
- V1-V600 (numerical variable): landmark coordinates in a 2d array format
Landmarks definition: landmarks and semilandmarks definition used for whole skull and each bone ((skull_slider_2.csv, angular_slide.csv, cer_slide.csv, dentary_slide.csv, hyo_slide.csv, nasal_slide.csv, neuro_slide_2.csv, pjaw_slide.csv, premax_slide.csv, uro_slide.csv)
- Before, Slide, Ater: positions and links of landmarks and semilandmarks
- Bones code: Angular (ang), Ceratohyal (cer), Dentary (dent), Hyomandibula (hyo), Nasal (nasal), Neurocranium (neuro), Pharyngeal jaw (pjaw), Premaxilla (premax), Urohyal (uro)
Labridae Tree (Labrid82ModTimeTree.tre): the phylogeny (TRE format)
- Phylogenetic tree with 412 tips and 411 internal nodes.
Classifiers (ontogeny_group.cvs): Table containing the name of the species indicating the classification according to two taxonomic groups
- Specimen (categorical variable): individual species' name
- Species (categorical variable): Currently valid scientific name
- Group (categorical variable): 1 = wrasses species; 2 = parrotfishes species;
- Group_2 (categorical variable): 1 = wrasses species; 2 = parrotfishes species; 3 = S. iseri individuals)
Classifiers (Scarus_LS.cvs): Table containing S. iseri individuals with total length (millimeters) and size class codes.
- Species (categorical variable): individual species' name and its identification number
- TL (numerical variable): total length (millimeters)
- Size_code (categorical variable): 1 = > 2.2 cm; 2 = 2.3 - 3.5 cm; 3 = 3.6 - 5.5 cm; 4 = 5.6 - 6.9 cm; 5 = 7.0 - 10.7 cm; 6 = 10.8 - 12.2 cm; 7 = 12.3 - 14.9 cm; 8 = 15.0 - 20.3 cm; 9 = 28.9 - 33.7 cm)
Classifiers (Scarins_group.cvs): Table containing all parrotfishes used in the study and their classifications between adults Scarines and S. iseri individuals from ontogeny series
- Species (categorical variable): Currently valid scientific name
- Group (categorical variable): Scarines (adult parrotfishes) or Scarus_ontogeny (S. iseri individuals from ontogeny series)
Classifiers (wrasses_group.cvs): Table containing all wrasses used in the study and their classifications between adults wrasses and S. iseri individuals from ontogeny series
- Species (categorical variable): Currently valid scientific name
- Group (categorical variable): Wrasses (adult wrasses species) or Scarus_ontogeny (S. iseri individuals from ontogeny series)
Code/Software
R is required to run Parrotfish_ontogeny.R; the script was created using version 4.3.2.\
Annotations are provided throughout the script through:
#STEP 1 - Scarus iseri ontogeny analysis
#STEP 2 - Comparing ontogenetic allometry of S. iseri to the evolutionary allometries of Scarinini and all Labrids
#STEP 3 - Comparing skull shape of S. iseri to the other Labrids
References
- Adams DC, Collyer ML, Kaliontzopoulou A, Baken EK. 2022 Geomorph: Software for geometric morphometric analyses. R package version 4.0.4. https://cran.rproject.org/package=geomorph
- Adams DC, Collyer ML. 2018 Phylogenetic ANOVA: group-clade aggregation, biological challenges, and a refined permutation procedure. Evolution 72, 1204–1215. (doi:10.1111/evo.13492)
- Collyer ML, Adams DC. 2018 RRPP: An R package for fitting linear models to high‐dimensional data using residual randomization. Methods Ecol. Evol. 9, 1772–1779. (doi: 10.1111/2041-210X.13029)
- Collyer ML, Adams DC. 2023 RRPP: Linear Model Evaluation with Randomized Residuals in a Permutation Procedure. R package version 1.4.0. https://CRAN.R-project.org/package=RRPP
- Evans KM, Larouche O, West JL, Gartner SM, Westneat MW. 2023 Burrowing constrains patterns of skull shape evolution in wrasses. Evol. Dev. 25, 73–84. (doi:10.1111/ede.12415)
- Gartner SM, Larouche O, Evans KM, Westneat MW. 2023 Evolutionary patterns of modularity in the linkage systems of the skull in wrasses and parrotfish. Integr. Org. Biol. 5, obad035. (doi:10.1093/iob/obad035)
- Larouche O, Gartner SM, Westneat MW, Evans KM. 2023 Mosaic evolution of the skull in Labrid fishes involves differences in both tempo and mode of morphological change. Syst. Biol. 72, 419–432. (doi: 10.1093/sysbio/syac061)
- Paradis E, Schliep K. 2019. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics,** 35**, 526–528. (doi:10.1093/bioinformatics/bty633)
- Pennell MW, Eastman JM, Slater GJ, Brown JW, Uyeda JC, FitzJohn RG, Alfaro ME, Harmon LJ. 2014. geiger v2. 0: an expanded suite of methods for fitting macroevolutionary models to phylogenetic trees. Bioinformatics 30, 2216–2218. (doi:10.1093/bioinformatics/btu181)
- Puttick M, Thomas G, Freckleton R, Clarke M, Ingram T, Orme D, Paradis E. 2019 motmot: Models of Trait Macroevolution on Trees. R package version 2.1.3. https://CRAN.Rproject.org/package=motmot
- R Core Team. 2023 R: A language and environment for statistical computing. Version 4.3.2. R Foundation for Statistical Computing
- Rhoda D, Polly PD, Raxworthy C, Segall M. 2021 Morphological integration and modularity in the hyperkinetic feeding system of aquatic-foraging snakes. Evolution 75, 56–72. (doi: 10.1111/evo.14130)
- Rhoda D, Segall M, Larouche O, Evans K, Angielczyk KD. 2021 Local superimpositions facilitate morphometric analysis of complex articulating structures. Integr. Comp. Biol. *61*, 1892–1904. (doi: 10.1093/icb/icab031)
- Zelditch ML, Swiderski DL. 2023 The predictable complexity of evolutionary allometry. Evol. Biol. *50*, 56–77. (doi: 10.1007/s11692-022-09581-1)
Methods
To investigate growth allometry, we examined the ontogenetic series of S. iseri, encompassing 54 individuals with a total length ranging from 1.75 to 33.5 cm.
For skull shape comparison analyses among S. iseri, other parrotfishes, and non-parrotfishes wrasses, our sample of adults comprises 336 individuals (160 adult parrotfishes and 176 adult non-parrotfish wrasses) from 217 labrid fishes (Table S1). From this dataset, we compared the evolutionary allometry of Scarines reef clade (57 species) to ontogenetic allometry of S. iseri. The ontogenetic series of S. iseri was obtained through sampling carried out in Belize in 2023 (IACUC protocols: UT Austin/AUP-2021-00064 and SERC/10-15-15-SJB; Collection permits: BF000005-16 and BF0042-22) with additional individuals from the Field Museum of Natural History (FMNH). The remaining species were obtained through several museums (Table S1).
We assessed the three-dimensional skull shape of each labrid species through micro-computed tomography (μCT) scans. Specimens were scanned at Rice University, the University of Texas Austin, the University of Minnesota, and the University of Washington Friday Harbor Labs in conjunction with the scanAllFishes and oVert initiatives. The scans were segmented in Amira v2.0.0 (Thermo Fisher Scientific, Waltham, MA) to isolate the skull while eliminating scales and debris. The isolated skulls were transformed into 3D meshes and imported into Checkpoint (Stratovan, Davis, CA). We followed the landmark scheme of [19,28], however, three additional landmarks detailing points along the pectoral girdle were incorporated [27]. The characterization of shape variation involved 200 3D points, 83 landmarks, and 117 semi-landmarks. These points comprehensively sampled the pharyngeal jaws, oral jaws, neurocranium, nasals, hyomandibula, operculum, hyoid apparatus, and pectoral girdle. All points were exclusively situated on the left side of the skull.
To accommodate for the rotation and translation inherent in highly kinetic articulating components of the fish skull), we initially aligned each dataset through Generalized Procrustes analysis (GPA) using the gpagen function. Subsequently, a local superimposition was conducted to standardize the positioning of diverse skull elements [38,39]. To analyze the shape evolution of individual bones, we partitioned our extensive skull shape dataset into smaller datasets for each bone, and the raw coordinates were individually superimposed. In this context, size was quantified as centroid size (CS), which represents the square root of the summed squared distances of each landmark to the centroid. After the local superimposition, we computed the average shapes and centroid sizes for the adult specimens of each species. All geometric morphometric analyses were carried out using the geomorph package, version 4.0.4 [40] in R, version 4.3.2 [41].