No evidence that the widespread environmental contaminant caffeine alters energy balance or stress responses in fish
Data files
Sep 01, 2023 version files 237.83 KB
-
caftimebin.csv
-
count.worms.eaten.csv
-
eat.csv
-
Fish.Lengths.Cleaned.csv
-
fish.measurements.foraging.csv
-
fish.measurements.full.csv
-
foragingdata.csv
-
o2.slope.csv
-
README.md
-
Tan_2023_activity_antipredator_behaviour.Rmd
-
Tan_2023_foraging_and_morph_data.Rmd
-
Tan_2023_Mosquitofish_metabolic_rate.Rmd
Abstract
Anthropogenic sources of environmental pollution are ever-increasing as urban areas expand and more chemical compounds are used in daily life. The stimulant caffeine is one of the most consumed chemical compounds worldwide, and as a result, has been detected as an environmental contaminant in all types of major water sources on all continents. Exposure of wildlife to environmental pollutants can disrupt the energy balance of these organisms, as restoration of homeostasis is prioritised. In turn, energy allocated to other key biological processes such as growth or reproduction may be affected, consequently reducing the overall fitness of an individual. Therefore, we aimed to investigate if long-term exposure to environmentally relevant concentrations of caffeine had any energetic consequences on wildlife. Specifically, we exposed wild eastern mosquitofish (Gambusia holbrooki) to one of three nominal concentrations of caffeine (0, 100, and 10,000 ng/L) and assayed individuals for metabolic rate, general activity, antipredator and foraging behaviour, and body size as measures of energy expenditure or energy intake. We found no differences in any measured traits between any of the given exposure treatments, indicating that exposure to caffeine at current environmental levels may not adversely affect the energy balance and fitness of vulnerable freshwater fish.
README: Title of Dataset: No evidence that the widespread environmental contaminant caffeine alters energy balance or stress responses in fish
Dataset is comprised of R scripts and .csv files that contain data used in the data analysis for the manuscript accepted at Ethology, doi: 10.1111/eth.13403, with the same title as the dataset.
Briefly, the activity, antipredator behaviour, foraging behaviour, metabolic rate, and morphological traits of eastern mosquitofish (Gambusia holbrooki) exposed to the environmental contaminant caffeine were analysed.
Description of the data and file structure
Contents of this dataset include 3 seperate R markdown (RMD) files for data analysis, and 8 .csv files containing data.
Titles of each R markdown file indicates which results section has been analysed.
Data from .csv files pertaining to each RMD file is included within the R script in each RMD.
The RMD file "Tan 2023 foraging and morph data" details the R scripts used to analyse the foraging behaviour and morphological traits of eastern mosquitofish exposed to caffeine.
Specifically, comparing caffeine exposure treatments (exposure), we analyse the latency to first consume a food item (time.to.first.eat) using a cox proportional hazards survival model, and the number of food items eaten (worms.eaten) using a zero-inflated poisson distributed generalised linear model.
Body mass (weight.g) and sex (sex.corr) are included as covariates in these models.
For morphological traits, body mass (weight.g) and total lengths (Total.Length) were analysed with analysis of variance models to determine differences between caffeine exposure treatment groups.
The RMD file "Tan 2023 activity antipredator behaviour" details the R scripts used to analyse the baseline activity and antipredator responses of mosquitofish exposed to caffeine.
Specifically, using linear mixed effect models, the distance travelled (distance.moved.cm) was compared between caffeine exposure treatment groups before mosquitofish were presented with a predatory stimulus as a measure of baseline activity, and after being presented with a predatory stimulus as a measure of antipredator response (time.bin).
Body mass (weight.g) and sex (sex.corr) are included as covariates in these models.
The RMD file "Tan 2023 Mosquitofish metabolic rate" details the R scripts used to analyse the oxygen uptake/metabolic rate of mosquitofish exposed to caffeine.
Specifically, the formula to calculate VO2 (i.e., non-mass specific rate of oxygen uptake) is presented in this file.
VO2 (vo2.mcg.h) was then analysed with linear mixed effect models to compare rates of oxygen uptake between caffeine exposure treatment groups (TREATMENT).
Body mass (weight.mg), sex (sex.correct), daily round of trial (ROUND), and activity (activity.cm) were included as covariates.
A random intercept for the 8 respirometry channels available (1|TANK.ID) was also included.
Code/Software
The open-source free software R (version used was 4.3.0) and RStudio is required to open RMD files and run through R scripts included in this dataset. R packages required to run analyses are detailed within each RMD file. Each RMD file also details which .csv files are required to be loaded.
Description of variables in CSV files
CSV File Name: fish.measurements.foraging.csv
Description of file:
Wet body weight data used as a covariate for models analysing mosquitofish foraging behaviour, see details in Tan et al. 2023 (doi: 10.1111/eth.13403).
Variables:
date: date of measurement
sex: sex of individual; JF and JM = juveniles (eventually consolidated into one group within analyses), M = males, F = females
weight.g: wet body weight of individuals in grams
exp.tank.id: experimental tank ID used for behavioural trial
exp.tank.temp: experimental tank temperature during behavioural trial in degrees celcius
Observation.id: unique observation ID assigned to each focal experimental individual (these are not tracked across different behavioural trials as animals are grouped housed, i.e., individuals with the same unique identifier are only specific to each type of behavioural trial).
exposure: caffeine exposure treatment group; c = control, l = low nominal treatment, h = high nominal treatment
CSV File Name: count.worms.eaten.csv
Description of file:
Data of number of food items consumed as one of the foraging behaviour endpoints measured in the foraging behaviour trial, see details in Tan et al. 2023 (doi: 10.1111/eth.13403).
Variables:
Observation.id: unique observation ID assigned to each focal experimental individual (these are not tracked across different behavioural trials as animals are grouped housed, i.e., individuals with the same unique identifier are only specific to each type of behavioural trial).
worms.eaten: number of food items (bloodworms) eaten across the length of the trial period
CSV File Name: caftimebin.csv
Description of file:
Activity data used within models analysing the general activity and antipredator behaviour of caffeine-exposed mosquitofish, see details in Tan et al. 2023 (doi: 10.1111/eth.13403).
Variables:
date: data of measurement
fish.id: unique observation ID assigned to each focal experimental individual (these are not tracked across different behavioural trials as animals are grouped housed, i.e., individuals with the same unique identifier are only specific to each type of behavioural trial).
exposure.tank.id: caffeine exposure treatment tank ID that focal individuals were housed in prior to experimentation
exposure: caffeine exposure treatment group; c = control, l = low nominal treatment, h = high nominal treatment
trial: trial ID used for identification purposes in ethovision tracking software
arena: arena ID used for identification purposes in ethovision tracking software (four arenas were present in each trial)
time.bin: time intervals trials were split into for analysis; Start-0:10:00 = initial 10 min acclimation period not included in analyses, 0:10:00-0:20:00 = 10 min baseline general activity period, 0:20:00-0:30:00 = 10 min post-predator strike activity
distance.moved.cm: distance travelled in centimeters (tracking software output)
mean.velocity.cm/s: mean velocity in centimeters per second (tracking software output)
cum.dur.in.arena.s: duration fish was tracked in seconds (tracking software output)
latency.to.arena.s: latency for fish to first enter arena (tracking software output)
sex: sex of individual; JF and JM = juveniles (eventually consolidated into one group within analyses), M = males, F = females
weight.g: wet body weight of individuals in grams
experimental.tank.id: experimental tank ID used for behavioural trial
experimental.tank.temp: experimental tank temperature during behavioural trial in degrees celcius
experimental.round: round of trial per day
CSV File Name: foragingdata.csv
Description of file:
Foraging zone latency data measured as one of the foraging behaviour endpoints in the foraging behaviour trial, see details in Tan et al. 2023 (doi: 10.1111/eth.13403).
Variables:
Observation id: unique observation ID assigned to each focal experimental individual (these are not tracked across different behavioural trials as animals are grouped housed, i.e., individuals with the same unique identifier are only specific to each type of behavioural trial).
Observation date: date of measurement
Total length: total length of trial video in seconds
acclimation period: trial video timestamp where acclimation period ends in seconds
first entry including acclimation period: first entry into foraging zone recorded in trial video
first entry: true first entry latency (first entry in trial video minus acclimation period timestamp)
CSV File Name: eat.csv
Description of file:
Foraging behaviour latency data measured as one of the foraging behaviour endpoints in the foraging behaviour trial, see details in Tan et al. 2023 (doi: 10.1111/eth.13403).
Variables:
Observation id: unique observation ID assigned to each focal experimental individual (these are not tracked across different behavioural trials as animals are grouped housed, i.e., individuals with the same unique identifier are only specific to each type of behavioural trial).
Observation date: date of measurement
Total length: total length of trial video in seconds
time to first eat including acclimation period: time to first consume a food item recorded in trial video (this includes the acclimation portion of the video and is not the true latency data)
CSV File Name: o2.slope.csv
Description of file:
Oxygen consumption data (% oxygen saturation over time) used to analyse metabolic rates of caffeine-exposed mosquitofish, see details in Tan et al. 2023 (doi: 10.1111/eth.13403).
Variables:
label: Arbitrary consecutive identifier
TRIAL ID: unique observation ID assigned to each focal experimental individual (these are not tracked across different behavioural trials as animals are grouped housed, i.e., individuals with the same unique identifier are only specific to each type of behavioural trial).
TREATMENT: caffeine exposure treatment group; C = control, L = low nominal treatment, H = high nominal treatment, B = blanks used to measure oxygen depletion in chambers not including a focal individual
DATE: date of measurement
TANK ID: experimental tank ID used in trial
ROUND: round of trial per day
Lbound: summary statistic provided by summary of function rankLocReg in R package LoLinR; left bound of time series local regression
Rbound: summary statistic provided by summary of function rankLocReg in R package LoLinR; right bound of time series local regression
alph: summary statistic provided by summary of function rankLocReg in R package LoLinR; window size used to fit local regression as proportion of data
b0: summary statistic provided by summary of function rankLocReg in R package LoLinR; y-axis intercept (% oxygen saturation) of local regression
b1: summary statistic provided by summary of function rankLocReg in R package LoLinR; slope of local regression (% oxygen saturation over time/every 30 seconds)
b1LoCI: summary statistic provided by summary of function rankLocReg in R package LoLinR; lower bound of 95% confidence interval for the slope (b1) of the local regression
b1UpCI: summary statistic provided by summary of function rankLocReg in R package LoLinR; upper bound of 95% confidence interval for the slope (b1) of the local regression
ciRange: summary statistic provided by summary of function rankLocReg in R package LoLinR; range of 95% confidence interval
skew: summary statistic provided by summary of function rankLocReg in R package LoLinR; skewness of standardized residuals for the local regression estimated as the Fisher-Pearson standardised third moment coefficient
bgN: summary statistic provided by summary of function rankLocReg in R package LoLinR; modified Breusch-Godfrey statistic used to estimate autocorrelation of standardised residuals
Lz: summary statistic provided by summary of function rankLocReg in R package LoLinR; combined linearity metric used to rank local regressions with implicitly weighting of each component metric by their empirical variances
Leq: summary statistic provided by summary of function rankLocReg in R package LoLinR; combined linearity metric used to rank local regressions with component metrics weighted equally
Lpc: summary statistic provided by summary of function rankLocReg in R package LoLinR; combined linearity metric used to rank local regressions with summed percentile-ranks of Zmin scores for each component metric. This ranking method was used in data analyses.
regressions.n: number of fitted local regressions per focal individual time series
SEX: sex of individual; JF and JM = juveniles (eventually consolidated into one group within analyses), M = males, F = females
sex.correct: sex of individuals with juveniles consolidated into one group; J = juveniles, M = males, F = females
WEIGHT: wet body weight of individuals in grams
chamber.vol: volume of respirometry chambers in milliliters
blanko2: slope of oxygen depletion of blank chambers used to account and control for background respiration, matched by experimental tank ID and day of trial
activity.cm: activity of individual within chamber measured as a subset of trial period to account for movement during oxygen consumption
CSV File Name: Fish.Lengths.Cleaned.csv
Description of file:
Standard and total lengths of mosquitofish used across all experimental trials, extracted from images
Variables:
Aim: Timepoint of measurement and corresponding experimental trial; Pre = pre-exposure prior to caffeine treatment, 1 = fish used in foraging and metabolic rate trials 21-days since start of caffeine exposure, 2 = fish used in activity and antipredator trials 42-days since start of caffeine exposure
Treatment: caffeine exposure treatment group; PE = pre-exposure, C = control, L = low nominal treatment, H = high nominal treatment
Treatment #: exposure tank ID
Fish Number: unique ID arbitrarily assigned to each fish within each exposure tank, treatment group, and timepoint
Date Taken: data photograph of fish was taken (following completion of respective experimental trial, or prior to exposure for pre-exposure fish)
Measurement: type of measurement taken; Standard.Length = standard length of fish in millimeters, Total.Length = total length of fish in millimeters
Order: arbitrary order of measurement taken
Length: length of corresponding measurement in millimeters
CSV File Name: fish.measurements.full.csv
Description of file:
Wet body weight of mosquitofish used across all experimental trials, with measurements taken following completion of respective experimental trial
Variables:
date: data of measurement
sex: sex of individual; JF and JM = juveniles (eventually consolidated into one group within analyses), M = males, F = females
weight.g: wet body weight in grams
exp.tank.id: focal individual's experimental tank ID within each experimental trial
Observation.id: unique ID assigned to each fish within each experimental trial
exposure: caffeine exposure treatment group; c = control, l = low nominal treatment, h = high nominal treatment, fwh and fwl = fish returned to freshwater from their respective low and high caffeine exposure treatments. However, note that these fish were not eventually used in analyses.
trial: type of experimental trial respective to each focal individual
timepoint: timepoint of measurement following start of caffeine exposure, 1 = 21 days since start, 2 = 42 days since start
Usage notes
R and RStudio