Authors: Hadis Miraly1,, N. Roxanna Razavi1,, Richard T. Kraus2, Ann Marie Gorman3, Elizabeth Duskey4, Matthew Altenritter5,†, and Karin E. Limburg1,6,*

Affiliations:

1State University of New York College of Environmental Science and Forestry, Syracuse, New York, United States

2U.S. Geological Survey, Great Lakes Science Center, Lake Erie Biological Station, Ohio, United States

3Fairport Fish Research Station, Ohio Department of Natural Resources, Fairport Harbor, Ohio, United States

4 Oklahoma State University, Department of Natural Resource Ecology and Management, Stillwater, Oklahoma, United States

5Department of Environmental Science and Ecology, State University of New York at Brockport, Brockport, New York, United States, †Deceased

6Department of Aquatic Resources, Swedish University of Agricultural Sciences, Uppsala, Sweden

*Corresponding authors: Hadis Miraly (hmiraly@esf.edu); N. Roxanna Razavi (razavi@esf.edu); Karin Limburg (klimburg@esf.edu)

Access this dataset on Dryad: doi:10.5061/dryad.vmcvdnd2b

Description of the data and files structure:

 Eye lens and otolith analyses

The data file named “Compiled lens-otolith data.csv” contains otolith Mn:Mg and eye lens Hg data for both Round Goby and Yellow Perch.

Compiled lens-otolith data.csv:

Otolith Mn:Mg Analysis

To examine differences in otolith Mn:Mg across basins and among age groups for each individual fish (for both Round Goby and Yellow Perch), linear mixed-effects regression models were applied separately for each species.

The R script used for this analysis is: otolitheyelenses-RGYP.R

Visualization of the results can be found in Figure 2 in the associated manuscript.

The code used to generate Figure 2 is also located in otolitheyelenses-RGYP.R.

Eye Lens Hg Analysis

To analyze differences in eye lens Hg levels across basins and among age groups for each individual fish (for both Round Goby and Yellow Perch), linear mixed-effects regression models were again applied separately for each species.

The R script used for this analysis is: otolitheyelenses-RGYP.R

Visualization of the results can be found in Figure 3 in the manuscript.

The code used to generate Figure 3 is included in the same script.

Total Length, Age Differences, and Muscle Tissue [Hg] Analyses

Length and Age Distribution

A one-way ANOVA was used to test for differences in total length between basins for each species.

A Chi-Square (χ²) test was conducted to evaluate whether the distribution of individuals across age groups differed between basins, by species.

Muscle Tissue [Hg] Analysis

To assess variation in muscle tissue [Hg] concentrations:

A two-way ANOVA was performed for each species to test for the effects of age and basin on muscle tissue [Hg].

The data file used: “MuscleHgTL.csv”

The corresponding R script: MuscleHg and TL.R

Conversion of Whole-Body [Hg] to Muscle Tissue [Hg] (Round Goby)

For Round Goby, small individuals (n = 41 of 74 total) were analyzed as whole-body samples for Hg. To ensure consistency with Yellow Perch data (which was based on muscle tissue [Hg]), a linear regression model was used to convert whole-body [Hg] to muscle tissue [Hg]. This model was developed using individuals for which both muscle and whole-body Hg concentrations were available.

The data file used: “MuscleHgTL.csv”

The corresponding R script: MuscleHg and TL.R

Visualization of the results can be found in Figure S1 in the supplementary information.

MuscleHgTL.csv:

RGwhole_tomusclehg.csv

System: Refers to the aquatic ecosystem sampled. In this case, "LE" stands for Lake Erie.

Site: Indicates the specific region within the system. "WB" represents the Western Basin of Lake Erie.

ID: Unique identifier assigned to each individual fish. For example, "RGWB-01" may stand for Round Goby from the Western Basin.

Age: The age of the fish in years.

Whole body Hg: The concentration of mercury in the whole body of the fish, typically measured in nanograms per gram (ng/g) dry weight.

Muscle Hg: The concentration of mercury in the muscle tissue of the fish, also likely in ng/g dry weight. Some values are missing (cells are blank), which indicate that muscle Hg was not measured Relationship Between Otolith Hypoxia Proxy and Eye Lens Hg Exposure

Exceedance of Hypoxia and Mercury Exposure Thresholds

To estimate the probability that each fish experienced high hypoxia exposure during its lifetime, we first established species-specific thresholds for both otolith Mn:Mg (as a hypoxia proxy) and eye lens Hg concentrations.

Threshold Calculation

For each species, a K-median clustering analysis was applied to partition Mn:Mg values along each fish’s lasered otolith transect into two distinct clusters:

One representing high hypoxia exposure

One representing low hypoxia exposure

The K-median method was chosen for its robustness to outliers, and the threshold was defined as the mean of the two cluster centroids. All data were centered and scaled prior to clustering.

Bootstrapping with 500 iterations (Leisch 2006) was used to assess the stability of the clustering results.

A similar K-median clustering approach was used to define species-specific thresholds for eye lens Hg concentrations.

The stability of both clustering procedures was confirmed using the mean adjusted Rand index, yielding values of 0.93 for Hg and 0.82 for Mn:Mg (Hubert and Arabie 1985).

Round Goby workflow

Step 1: Data Preparation

• Reorganized the dataset, added basin information and normalized distances
• Generated .txt files (eyelens_goby.txt and otolith_goby.txt ) for both otolith and eye lens data using goby_data.R
• Data files used: “goby_all.xlsx” and “goby_id.csv”
• R script: “goby_data.R”

Step 2: Threshold and Probability Calculation

• Calculated hypoxia and Hg thresholds
• Computed the conditional probability that each fish experienced high hypoxia exposure
• R script: “goby_Hypoxia-Hg.R"

Yellow perch workflow

Step 1: Data Preparation

• Reorganized the dataset, added basin information and normalized distances
• Generated .txt files (eyelens_perch.txt and otolith_perch.txt) for both otolith and eye lens data using “perch_data.R”

eyelens_goby, eyelens_perch and otolith_goby, otolith_perch txt files

• Distance: Distance from a reference point, likely the core of the otolith or lens (in micrometers or arbitrary units).
• MnCa: Ratio of Manganese to Calcium; a chemical signature used for environmental or physiological analysis (present only in eyelens files).
• MnMg: Ratio of Manganese to Magnesium; another geochemical marker.
• Age: Estimated age
• ID: Sample or specimen identifier.
• Basin: Basin code or name where the specimen was collected (e.g., "WB" for White Basin, "CB" for Clear Basin).
• Dnorm: Normalized distance; the distance value scaled between 0 and 1.
• MnMg_scale: Scaled MnMg ratio, likely standardized for analysis (e.g., z-score or centered).

• Data files used: “perch_all.xlsx” and “perch_id.csv”
• R script: “perch_data.R”

Step 2: Threshold and Probability Calculation

• Calculated hypoxia and Hg thresholds
• Computed the conditional probability that each fish experienced high hypoxia exposure
• R script: “perch_Hypoxia-Hg.R”

goby_all.xlsx and perch_all.xlsx:

goby_id.csv and perch_id.csv:

 R packages

ggplot2

ggpubr

tidyr

dplyr

gridExtra

lme4

lmerTest

tidyverse

emmeans

ggplot2

ggpubr

ggridges

tibble

patchwork

scales

flexclust

References:

Hubert, L. & Arabie, P. (1985). Comparing partitions. Journal of Classification, 2, 193-218. 10.1007/BF01908075

Leisch, F. (2006). A toolbox for K-centroids cluster analysis. Computational Statistics & Data Analysis, 51, 526-544. 10.1016/j.csda.2005.10.006