Compounding heterochrony shapes the salamander visual system across adaptive zones
Data files
Aug 19, 2025 version files 31.59 MB
-
1000_Chronograms_559_Salamander_lineages.tre
31.54 MB
-
BAYESTRAITS_COMMANDS_12April2025.rtf
1.51 KB
-
Discrete_Photoreceptor_dataset_18Aug2025.txt
17.16 KB
-
Eye_Retina_dataset_18Aug2025.csv
16.27 KB
-
OUwie_COMMANDS_6Aug2025.R
2.76 KB
-
PGLS_and__Phylolm_COMMANDS_12April2025
8.04 KB
-
README.md
6.72 KB
Abstract
Transitioning between disparate environments presents new physical challenges, and metamorphosis can provide solutions. The life cycles of most amphibians involve an aquatic-to-terrestrial transition and concomitant metamorphosis, but shifts in developmental timing (heterochrony) have also produced a wide variety of aquatic-only and terrestrial-only forms. Thyroid hormone signaling governs the timing of tissue transformation and may be a key mechanism behind the relationship between development and diversification of some metamorphic traits. Here, we show that life-cycle mode and cave-adaptation (troglomorphy) through heterochrony primarily explain variation in salamander eyes and retina. Across levels of organization (organ/tissue), heterochrony led to serial reductions in the visual system of larval-form paedomorphs that lost metamorphosis. This pattern is compounded in larval-form paedomorphic lineages that subsequently transitioned to subterranean environments and evolved troglomorphic traits. Following Haller’s Rule, visual system investment declines across ontogeny but at a faster rate in paedomorphs and even faster (for eye size) in troglomorphs. Thyroid hormone typically increases eye size during metamorphosis, however, we show responsiveness is reduced in paedomorphs and lost or reversed in troglomorphs. Salamander visual system variation is an example of how alterations to hormone-mediated transformation can shift developmental trajectories and compound phenotypic modifications as species move into more specialized environments.
https://doi.org/10.5061/dryad.8931zcs04
This database includes:
- one file with 1000 phylogenetic trees (chronograms). Life cycle and habitat/ecological data were compiled from the literature. Morphometric data are primarily external measurements. Retinal cell count data were from histological slides and published images. Posterior distribution of 1000 trees with 559 taxa was reconstructed in BEAST.
- two data files containing the values used in phylogenetic comparative analyses
- three files with base commands for BayesTraits, OUwie, PGLS, and Phylolm
*See Manuscript and Electronic Supplemental Material for methodological details, alternative regime models, and specimen numbers.
TREES FILE: “1000_Chronograms_559_Salamander_lineages.tre”
A file with 1000 post-burnin Bayesian Chronograms of 559 salamander lineages.
Names match those listed in the data files.
DATA FILE: “Eye_Retina_dataset_18Aug2025.csv”**
Species: Lists 156 species (lineages) included in one or more of the analyses.
LifeCycle is coded as three states: pd = obligate paedomorphic, bi = biphasic, dd = direct development
Eco_LJ is the habitat/ecology of the larval/juvenile stage in 4 categories: aquatic (lo = lotic or le = lentic) and terrestrial (gd = ground-dwelling or cb = climbing)
Eco_A is the habitat/ecology of the adult stage in 4 categories: aquatic (lo = lotic or le = lentic) and terrestrial (gd = ground-dwelling or cb = climbing)
Eco_SfTg is surface or “obligate” cave-dwelling as 2 categories: sf = surface-dwelling or tg = troglomorphic (“obligate” cave-dwelling, but we acknowledge local exceptions).
EWHL_Stg40 is the relative eye width at Harrison Embryonic stage 40. Eye width is scaled to head length.
EWHW_H is the relative eye width at hatching regardless of embryonic staging. Eye width is scaled to head width.
EWHW_LJ is the relative eye width at the larval/juvenile stage. Eye width is scaled to head width.
EWHW_A is the relative eye width at the adult stage. Eye width is scaled to head width.
RGC_dep_LJ is the depth of the RGC (Retinal Ganglion Cell) layer based on cell counts for the larval/juvenile stage.
INL_dep_LJ is the depth of the INL (Inner Nuclear Layer) based on cell counts for the larval/juvenile stage.
ONL_dep_LJ is the depth of the ONL (Outer Nuclear Layer) based on cell counts for the larval/juvenile stage.
Total_dep_LJ is the total cell depth (RGC + INL + ONL) based on cell counts for the larval/juvenile stage.
GC_dep_A is the depth of the RGC (Retinal Ganglion Cell) layer based on cell counts for the adult stage.
INL_dep_A is the depth of the INL (Inner Nuclear Layer) based on cell counts for the adult stage.
ONL_dep_A is the depth of the ONL (Outer Nuclear Layer) based on cell counts for the adult stage.
Total_dep_A is the total cell depth (RGC + INL + ONL) based on cell counts for the adult stage.
RGC_CompSectHW_LJ is the RGC count of a complete section for the larval/juvenile stage, scaled to head width.
INL_CmpSectHW_LJ is the INL cell count of a complete section for the larval/juvenile stage, scaled to head width.
ONL_CompSectHW_LJ is the ONL cell count of a complete section for the larval/juvenile stage, scaled to head width.
Total_CompSectHW_LJ is the total (RCG + INL + ONL) cell count of a complete section for the larval/juvenile stage, scaled to head width.
RGC_CompSectHW_A is the RGC count of a complete section for the adult stage, scaled to head width.
INL_CmpSectHW_A is the INL cell count of a complete section for the adult stage, scaled to head width.
ONL_CompSectHW_A is the ONL cell count of a complete section for the adult stage, scaled to head width.
Total_CompSectHW_A is the total (RCG + INL + ONL) cell count of a complete section for the adult stage, scaled to head width.
Cell_Sz is the cell size based on two 2D INL cell measurements
Gnm_Sz is the genome size based on c-values from the literature
GillL_Chng_TH is the measured percent change in gill length with 5 nM TH treatment for 21 days
EW_Chng_TH is the measured percent change in eye width with 5 nM TH treatment for 21 days
" - " Indicates no data
DATA FILE: “Discrete_Photoreceptor_dataset_18Aug2025.txt” (tab delimited)
Species: Lists 559 species (lineages) included in one or more of the multistate and discrete analyses.
LifeCycle is coded as three states: pd = obligate paedomorphic, bi = biphasic, dd = direct development
Eco_SfTg is surface or “obligate” cave-dwelling as 2 categories: 1 = surface-dwelling or 0 = troglomorphic (“obligate” cave-dwelling, but we acknowledge local exceptions).
PHOTO_LJ is whether the photoreceptor segment is present 1 or absent 0 for Larvae/Juveniles
PHOTO_A is whether the photoreceptor segment is present 1 or absent 0 for Adults
Data is only included when observed. Otherwise, we used " - ", which is a missing data symbol for BayesTraits.
" - " Indicates no data
**ANALYSIS FILE: “BAYESTRAITS_COMMANDS_12April2025.rtf” - Run in Mac Terminal
BayesTraits v4.1.1 was used to reconstruct ancestral life cycle modes (pd, bi, dd) and ecology (sf, tg) using an exponential prior set to 0.0005.
It was also used to test for associations (Dependencies) between binary traits. Specifically, we tested whether troglomorphy is dependent on paedomorphosis and the loss of the photoreceptive segment (PHOTO) in larvae/juveniles and adults.
**ANALYSIS FILE: “OUwie_COMMANDS_6Aug2025.R” - Run in R
OUwie v 2.10 was used to compare the fit of various life cycle and ecological models to continuous variables (relative eye size and various retinal cell depth measurements).
This approach allows for testing hypotheses of evolutionary shifts in the optimum (θ) and rate of evolution (σ2) of a continuous trait in response to different categorical selective regimes, such as life cycle mode (paedomorphic, biphasic, direct development) and/or habitat (obligate cave, aquatic, terrestrial). Models are defined in the Electronic Supplemental Materials document.
**ANALYSIS FILE: “PGLS_and__Phylolm_COMMANDS_12April2025” - Run in R
We used Phylogenetic Generalized Least Squares (PGLS) implemented in the R package caper to determine whether genome size (c-value) is correlated with retinal cell size (based on 2D measurements of INL soma), which could impact retinal cell parameters. Also included in the file are base codes for Phylogenetic Linear Regressions, which we used to test for differences in life cycle mode or habitat/ecology on retinal variables (RGC and INL depth) while accounting for covariates, genome size, or cell size.
This database includes two files with data and one file with trees. Life cycle and habitat/ecological data were compiled from the literature and personal observations. Morphometric data are primarily external measurements. Retinal cell count data were from histological slides and published images. Names and codes for each variable are described in the README file. Posterior distribution of 1000 trees of 559 taxa was reconstructed in BEAST. See the associated manuscript and electronic supplemental document for additional methodological details, alternative regime models, and specimen numbers.