McWhinnie et al. - Temperature and mass scaling affect cutaneous and pulmonary respiratory performance in a diving frog
Ryan B McWhinnie: [email protected]
Jason P Sckrabulis: [email protected]
Thomas R Raffel: [email protected]
Conceptualization: TRR, RBM, & JPS. Methodology: TRR, RBM, & JPS. Analysis: RBM, TRR, & JPS. Respirometer design: JPS. Respirometer construction: JPS & RBM. Writing – review & editing: RBM & TRR. Funding acquisition: TRR. Literature searching: RBM & TRR.
Please email [email protected] with any issues or suggestions for the respirometry device, or submit via the GitHub issues tab for this repository. For questions concerning the manuscript, please email the corresponding author at [email protected].
- Jan 20, 2022: Added full citation from Integrative Zoology.
- Apr 22, 2021: Upload .csv datasets & .R specialty code. Added variable descriptions and air density variables to
README.md. - Aug 12, 2019: First full commit.
Please cite work as:
McWhinnie RB, Sckrabulis JP, and TR Raffel. (2021) Temperature and mass scaling affect cutaneous and pulmonary respiratory performance in a diving frog. Integrative Zoology 16(5): 712-728. https://doi.org/10.1111/1749-4877.12551.
Our goal for this project was to develop a low-cost alternative to conventional commerical respirometry apparatuses in order to conduct temperature experiments with true replication of temperatures, which requires measuring many animals at the same time. As a result, we developed a flow-through respirometry device able to measure flow rate (SCCM) and percent oxygen (volumetric) simultaneously. Our device is based on the Arduino platform and uses components found at a majority of electronics suppliers (i.e., Mouser and Adafruit). Each device is under $200 (USD) with a footprint smaller than a mousepad at the time of purchase (2018). During our experiments, we were able to describe multiple aspects of respiratory activity in the diving frog Xenopus laevis. See full manuscript at the above citation for this analysis.
- README.md
- LICENSE.txt
Bill of Materials.csv
Parts list, excluding an enclosure, O2 modification, general supplies & equipment, and microSD card- data
Folder of experimental data as .csv files used in data analysis for manuscript (variable descriptions are below)
Example respirometry measurement.csvMcWhinnie breath bout data.csvMcWhinnie master data.csv
- code
Folder of specialized R code (as .R files) used for data processing and respirometer Arduino code (as .ino file)Individual breath bout oxygen calculation.RTotal pulmonary oxygen calculation.RBlock oxygen integrals.RRespirometryDevice.ino
- imgs
Folder of images used inREADME.md
For our respirometry devices to be viable, low-cost alternatives to commerical products, we opted to develop the device on the Arduino platform. Arduino is an open-source microcontroller platform designed to be accessible to new users with little programming and/or electronics expertise. Since its release, a mass accumulation of documentation, tutorials, and code examples exists on the internet. Due to its open-source nature, other companies have taken the Arduino platform and tailored it to specific needs (e.g., Adafruit). Adafruit's Pro Trinket product line further reduced costs by reducing the number of available analog and digitial I/O pins, footprint. For our fast processing and 5V sensor requirements, we utilized the Adafruit Pro Trinket 5V 16MHz microcontroller (Fig. 1A, #1) for our devices. In our case, we used right-angle headers for the FTDI pins of the Pro Trinket to better fit our enclosure and allow for flow rate calibration when connected to a computer.
We used the Honeywell ZephyrTM analog airflow rate sensor (Fig. 1A, #5; Fig. 1C, #9) to measure flow in standard cubic centimeters per minute (SCCM). This analog version of the sensor directly measures resistance change of temperature-sensitive resistors influenced by the flow rate. This line of sensors was designed for medical and industrial use and has a fast response time, making it ideal to capture small changes in flow produced by the lung movements of air in small frogs and other amphibians. The ZephyrTM line also has a linear output and customizable packages for easier integration by the end user's application.
We used the Seeed Studio Grove oxygen sensor package (Fig. 1A, #4) designed for Arduino applications. This package uses a replacable MEQ-O2 gas sensor, designed for use in enclosed areas (i.e., mines). Oxygen was reported as a volumetric percentage of air by measuring current change based on the electrochemical reaction of the gas on a heated electrode. We modified the sensor (Fig. 1C, #8) for use in a flow-through system by modifying and sealing a plastic coin case and air tubing connectors. Within the coin case, we made a channel with cut nylon washers and hot glue such that only the supplied air for respiration was accessed by the sensor.
For experimental use, we added a microSD card (Fig. 1A, #7) and real-time clock modules (Fig. 1A, #6). This allowed for data collection throughout the entire respiration period (~90 minutes in our experiments). For ease of use, the device has an integrated power jack, which can by powered by an appropriate DC power supply adapter, or 9V battery power (with connector) (Fig. 1A, #2). The device supports up to 100 unique data files (microSD card capacity dependent), as well as be controlled via switch (Fig. 1A, #3), and indicator LEDs (Fig. 1A, #2) for status for further ease of use. Completed device photos are presented in Fig. 1B & C.
The respirometry device was designed for ease of use. See Figure 2 for a pictoral representation of the following text, but see RespirometerDevice.txt for complete, commented operational code. Upon receiving power, the device checks for a valid microSD card, and enters Standby until the switch is flipped. When the switch is flipped and the digital input is detected, the device creates and opens a new data log file and enters Warm-up, where voltage is supplied to the oxygen sensor for 20 minutes followed by 30ms for the flow rate sensor (manufacturer specifications). After Warm-up, the device enters Collection and records the actual start time as tracked by the RTC. It is important to note that RTC time regularly drifts, and we recommend initializing RTC prior to every experimental block. In Collection the device measures voltage and current change of both sensors, calculates flow rate and oxygen percentage, and loads them into memory. When the user flips the switch again, the device records all data to the data file, closes it, and enters Standby. The device supports multiple experimental periods by continuing to enter and exit Standby following measurements. We were able to collect and average of ~80 measurements per second, allowing us to measure small changes in flow rate by lung movement of small frogs and amphibians.
Measurements logged by our respirometry device for a single animal's measurement are provided in data. Once an individual's respiratory performance was measured, we quantified metabolism as four proxies: 1) cutaneous respiration, 2) pulmonary respiration, 3) total respiration, and 4) breath rate. For a detailed description of experimental methods, see the full manuscript. Our device was used to quantify only 2 & 4 though direct calculation of oxygen (2) and counting of individual breaths (4). All analyses and data manipulation were done in R (v3.5.1; https://www.r-project.org).
To quantify pulmonary respiration, we needed to establish a baseline level for flow rate and percent oxygen, as defined by the values for each of these parameters when frogs were not breathing. Diving frogs like X. laevis typically have extended 'gap' periods with no breaths punctuated by distinct periods of breathing activity, which will be referred to from here on as 'breath bouts'. When plotted as a time series, these breath bouts are visually distinguishable from the gaps between breaths, making it possible to select representative baseline data between breath bouts and use it to establish a continuous running baseline through the entire time series. We used the fhs function from the gatepoints package to freehand select representative data for each baseline, excluding data more than 5 mL min−1 or 0.1% away from the visually apparent baseline for each flow or O2 dataset, respectively (Fig. 3). Once this was done, the na.fill function from the zoo package was used to fill 'NA's from the selected baseline dataset based on the surrounding data. To finalize baseline correction, a cubic smoothing spline was fit to each baseline dataset to generate a continuous baseline function (smooth.spline from the base package). The spline fit was then subtracted from the original (raw) dataset to generate a baseline-corrected data series centered around zero for flow rate (e.g., Fig. 3B) and percent oxygen (e.g., Fig. 3D). Following baseline correction, breath bout volumes and total oxygen consumption were calculated by multiplying the change in flow rate or percent oxygen from the baseline against the time frame for each specific measurement (i.e., the integral of the curve). These values yield the volume of flow or percent oxygen for each reading with sums being added for both positive and negative values to determine respective values for the entire measurement. See full manuscript for actual calculations.
The following information was used to calculate air density for July and August 2018 using information from WUnderground. See Supplemental Material of MS for full description of this calculation.
| Variable | July Value | August Value |
|---|---|---|
| Elevation (m) | 286 | 286 |
| Room Air Temp. (C) | 20 | 20 |
| Altimeter Setting (Inch-Hg) | 30 | 30 |
| Dew Point (F) | 58 | 62 |
| Air pressure (atm) | 1.002663 | 1.002663 |
Relative Humidity based on August-Roche-Magnus approximation
McWhinnie master data.csv
| Variable Name | Description |
|---|---|
| FrogID | Identification of individual frog |
| MeasurementID | Identification of a given measurement for a given frog |
| Block | Temporal block the animal was measured in |
| DayNum | One of five days performance measurements were taken on (-1, 0, 1, 4, 8) (For the additional post-experimental measurements, DayNum = 0 is used as a place holder since there was no acclimation period for these frogs) |
| Day | One of six performance measurements (distinguishes between performance measurements one and two for the zero-day measurements) |
| AccMassChange(g) | Change in mass in grams of a given frog during the acclimation period |
| AvgMass(g) | Average of mass of each frog taken immediately before and after the acclimation period in grams |
| AccTemp | Temperature of acclimation in Celsius |
| AccInc | Incubator housed in during acclimation period |
| PerfTemp | Temperature at which metabolic performance was measured at in Celsius |
| PerfInc | Incubator number housed in during performance measurement |
| BreathCount | Number of exhales plus inhales recorded from the animal during the measurement period |
| Length(s) | Length of measurement in seconds |
| Length(min) | Length of measurement in minutes |
| Length(hr) | Length of measurement in hours |
| BreathsPerHr | Number of exhales plus inhales recorded from the animal during the measurement period divided by measurement length in hours (BreathCount/Length(hr)). |
| BotHeight(cm) | Height of bottle in centimeters used to enclose the animal during performance measurements |
| TotWatVol(L) | Total volume of water the animal was in during performance measurements in liters |
| InitialWatDO(mg/L) | Amount of dissolved oxygen that was recorded just before animal was placed in enclosure for the performance measurement in milligrams O2 per liter |
| InitialWatDO(mg) | Amount of dissolved oxygen, in milligrams O2, that was recorded just before animal was placed in enclosure for the performance measurement in millgrams O2 calculated by multiplying InitialWatDO(mg/L) by TotWatVol(L) |
| InitialWatTemp | Temperature of the water that was recorded just before animal was placed in enclosure for the performance measurement in Celsius |
| InWatVol(L) | Total volume of water inside the bottle the animal was in during performance measurements in liters |
| InWatDO(mg/L) | Amount of dissolved oxygen inside the bottle that was recorded just before animal was placed in enclosure for the performance measurement in milligrams O2 per liter |
| InWatDO(mg) | Amount of dissolved oxygen inside the bottle, in milligrams O2, that was recorded just before animal was placed in enclosure for the performance measurement in milligrams O2 calculated by multiplying InWatDO(mg/L) by InWatVol(L) |
| InWatTemp | Temperature of the water inside the bottle that was recorded just before animal was placed in enclosure for the performance measurement in Celsius |
| OutWatVol(L) | Total volume of water outside the bottle the animal was in during performance measurements in liters |
| OutWatDO(mg/L) | Amount of dissolved oxygen outside the bottle that was recorded just before animal was placed in enclosure for the performance measurement in milligrams O2 per liter |
| OutWatDO(mg) | Amount of dissolved oxygen outside the bottle that was recorded just before animal was placed in enclosure for the performance measurement in milligrams O2 |
| OutWatTemp | Temperature of the water outside the bottle that was recorded just before animal was placed in enclosure for the performance measurement in Celsius |
| TotCutO2Cons(mg) | Total oxygen consumed cutaneously throughout the performance period (InWatDO(mg)+OutWatDO(mg)). |
| CutO2PerfRate(mg/hr) | Cutaneous rate of oxygen consumed per hour calculated by dividing TotCutO2Cons(mg) by Length(hr) in milligrams O2 per hour |
| CutMircomolO2/gFrog/hr | Total oxygen consumed cutaneously throughout the performance period calculated by dividing CutO2PerfRate(mg/hr) by AvgMass(g) in milligrams O2 per grams of frog per hour |
| PulmO2Cons(mg) | Total oxygen that taken up through the lungs throughout the performance period in milligrams O2 |
| PulmO2PerfRate(mg/hr) | Pulmonary rate of oxygen consumed in milligrams O2 per hour |
| PulmMicromolO2/gFrog/hr | Total oxygen taken up through the lungs throughout the performance period calculated by dividing PulmO2PerfRate(mg/hr) by AvgMass(g) in milligrams O2 per gram frog per hour |
| TotO2Consumed(mg) | Total oxygen that was consumed via both cutaneous and pulmonary means (TotCutO2Cons(mg)+PulmO2Cons(mg)) |
| TotO2PerfRate(mg/hr) | Total rate of oxygen consumed calculated by dividing TotO2Consumed(mg) by Length(hr) in milligrams per hour |
| TotMicromolO2/gFrog/hr | Total oxygen consumed via both cutaneous and pulmonary means calculated by dividing TotO2PerfRate(mg/hr) by AvgMass(g) in milligrams O2 per gram frog per hour |
| PropCut | Proportion of total O2 consumption performed cutaneously |
| PropPulm | Proportion of total O2 consumption performed via pulmonary means |
McWhinnie breath bout data.csv
Variables with the same name are identical to those above, except where noted below.
| Variable Name | Description |
|---|---|
| MeasurementID | For the breath bouts, animals' performance was measured at one of two sequential time points (ZeroOne & ZeroTwo) at zero days post acclimation |
| Breaths | Total number of measured inhales and exhales for a given breath bout |
| BreathsPerMin | Number of total breaths taken during the breath bout per minute |
| mgO2ConsPerBreath | Average mass of oxygen consumed in milligrams O2 per breath |
| mLAirPerBreath | Average volume of air breathed in milliliters air per breath |
Example respirometry measurement.csv
Calculations for these values can be found in RespirometryDevice.ino.
| Variable Name | Description |
|---|---|
| Date | Measurement start date calculated by real-time clock module |
| ReadTime | Time of day of measurement start date calculated by real-time clock module |
| Millis | Milliseconds since power was provided to device for time series |
| FlowAnalog | Number of analog reads (0 to 1023) on the flow rate sensor pin |
| FlowVoltage | Calculated voltage based on FlowAnalog |
| FlowRate | Calculated flow rate in SCCM based on FlowVoltage |
| O2Analog | Number of analog reads (0 to 1023) on the oxygen sensor pin |
| O2Voltage | Calculated voltage based on O2Analog |
| O2Percent | Calculated oxygen percentage based on O2Voltage |