CalR Updates and Known Bugs
CalR Version 2 Updates
Additions:
- Plotting can now be done accorrding to the highest
frequency of data collected.
- Timeplots can now be vizualized with a smoothed rolling average.
- Quality control plot has been added to plot mass change vs
total energy balance.
- Added the option to manually add mass change for quality
control plots.
- Reworked various functions to perform more efficiently such
as timeplots and remove outliers.
- Export hourly data to GraphPad Prism files
- Added power calculations to help vizualize needed sample sizes.
- Energy Balance has been added to statistical analysis.
- Timeplots ui tools that will help improve vizualization.
- Changelog tab was added which details frequent updates.
- C13 data from isotope analyzers from a "Sable" system can
be read into CalR for plotting.
- Enviornmental columns from "Sable" systems can be read into
CalR for plotting.
- Added two different methods of combining runs: continuous
and split.
Changes:
- Bar plots have been replaced by box plots
- Improved speed and performance by plotting with Javascript.
- Regression plots display kcal/hr instead of kcal/period.
CalR Version 1.3 Updates
New Features:
Cumulative Energy Expenditure
And
Cumulative Energy Balance
Small differences in food intake or energy expenditure over time can accumulate into changes in body weight. However due to small effect sizes and the variability between animals the changes in food intake or energy expenditure often do not reach statistical significance. By showing cumulative changes in energy balance over time, we can extract more meaning from each experiment.
In CalR version 1.2 we introduced the energy balance visualization. Energy balance uses the food intake in kcal (the caloric content of the food multiplied by mass of food consumed) and subtracts the energy expenditure over that time period. Now in Legacy CalR.3 we introduce the cumulative energy balance. We also now provide cumulative energy expenditure as a useful comparator to cumulative food intake.
Please see below for examples:
Known & persisting issues:
- System: CLAMS
- If recording start on exactly 00 seconds then it is possible that an error in CalR reading the time stamps may occur.
- System: TSE
- Some files appear to report cumulative food intake instead of instantaneous food intake. CalR does not handle this situation well.
- TSE files that contain semicolons in their name will be disconnected from the server.
- TSE generated files may contain a "Start Time" row that will cause the file to disconnect from the server.
Please report any new issues to us at wbernard@bidmc.harvard.edu
Bug Fixes:
- Fixed multiple occasions where light cycles were displayed incorrectly.
- Fixed spacing for analysis page where analysis titles would overlap analysis output.
- Fixed bug where bar plot y-axis labels would overlap with plot ticks.
- Fixed bug where bar plots would load twice before being displayed.
- Fixed Energy Balance bar plots.
- Fixed fence post error in time series selection
- Many users select hours 0-24 assuming that they have selected one day (24 hours) when actually all 25 hours were selected and displayed as hour 0 is included. Previously only the “Bar Plots” excluded the final selected hour (analysis of 24 hours), but other tabs (regression plots, analysis plots) included all 25 hours in the analysis. We have extended the removal of the last hour selected to all functions in CalR to appropriately reflect what the user believes they selected. Please note that to replicate analysis performed under Legacy CalR.2, the user must extend the time selected by an extra hour.
New features:
Increased speed of analysis following code optimization.
Time series plots:
- Added cumulative energy expenditure and cumulative energy balance.
- Added check for xytot or locomotor activity to distinguish if it is cumulative or not to allow files from older systems to be analyzed correctly.
- Added ability for wheel counts to be read in from TSE files.
- The hourly differences plot and the de-trended plot button have been deprecated.
- Added mass plotting based off of continuous mass recordings in TSE and Sable files.
Average Plots:
- “Average plots” was changed to “Bar Plots” to better represent what is may be displayed as not all plots before were technically averaged bar plots (e.g. cumulative values, not averages).
- Added variable selection and plot button to allow users to perform selection on the bar plots tab independently of the variable selected on the time series plots.
- Added more “info buttons” to briefly explain how plots are generated and what is being shown.
- Changed plotting from interactive to static plotting due to computational speed complaints.
- Removed selection of cumulative bar plots and instead overall cumulative plots are displayed in addition to average plots for variables that are derived from or already used to calculate cumulative variables. (food, drink, EE, EB, Wheel)
- Added Titles to all bar plots as well as more informative axis labels.
Regression Plots:
- Removed Energy Balance from regression plot selection.
- Please see analysis page changes for more in-depth explanation.
- Changed the unit scale of some regression plots from per hour to per phase period to better align with new bar plotting.
- Added more informative plot labels
Analysis Page:
- Removed cumulative GLM analysis as they were incorrect.
- Previous versions of CalR displayed p-values not derived from the max values of the cumulative variable.
- Removed Energy Balance from GLM analysis.
- We’ve decided that we need further study to understand how to analyze statistical significance of energy balance. Our initial approach followed this reasoning: for values that depend on mass (e.g. energy expenditure) we use GLM. For mass-independent values (e.g. RER) we use ANOVA. Since energy balance is calculated as the difference of two mass-dependent variables (food intake minus energy expenditure) energy balance therefore should also be a candidate to be analyzed via GLM. Yet so far our analysis of experiments has not shown a relationship between energy balance and mass suggesting that it might be more appropriate for analysis by ANOVA. Until this is resolved, CalR does not report statistical significance values for energy balance. We look forward to hearing input from the user community.
Anticipated features in the next versions of CalR
- Plotting with greater frequency than one point per hour
- Improved speed and performance
- Export summary data to GraphPad Prism files
- Time series spectral analysis & metabolic flexibility
- Estimating statistical power
- Making CalR source code available to the community on GitHub
- Making available additional R code for plotting data and analyzing CalR files
- Pilot version of a database for indirect calorimetry experiments using CalR files
- Support for SI units, from kcal to kJ