plethR is an R package for analyzing whole body plethysmography data from DSI systems
What it does: - Imports multi-sheet Excel files - Organizes subjects into experimental groups - Calculates group averages and statistics - Creates publication-quality figures - Performs PCA and clustering analysis
Why use it? - Standardized workflow - Reproducible analysis - Saves hours of manual work
Total: ~120 minutes
Study Design: - Groups: Uninfected WT, Uninfected Benac, Infected WT, Infected Benac - N per group: 4 mice - Total: 16 mice - Time series: Multiple measurements/week over multiple weeks - Parameters: f, TVb, MVb, Penh, PAU, and more
The original, simpler infection study
# Install devtools if you don't have it
install.packages("devtools")
# Install plethR from GitHub
devtools::install_github("camronp/plethR")
# Load the package
library(plethR)First time only! After installation, just use
library(plethR) each session.
# Load package
library(plethR)
# Check package help
?plethR
# List all functions
ls("package:plethR")
# Check a specific function
?sheets_into_listTroubleshooting: If you get errors, make sure all dependencies installed correctly.
# Create project folder structure
dir.create("WBP_Analysis", showWarnings = FALSE)
dir.create("WBP_Analysis/data", showWarnings = FALSE)
dir.create("WBP_Analysis/figures", showWarnings = FALSE)
dir.create("WBP_Analysis/results", showWarnings = FALSE)
# Set working directory
setwd("WBP_Analysis")
# Check
getwd()Best practice: Always use a project folder with organized subfolders!
DSI Excel file structure: - Each sheet = one mouse’s time series data - Columns = respiratory parameters (f, TVb, Penh, etc.) - Rows = measurements over time - May include “Apnea” sheets (usually not needed)
Common issues: - Date/time formats vary - Multiple measurements per day - Metadata columns we don’t need
# Basic import
df_list <- sheets_into_list("data/my_experiment.xlsx")
# With cleaning options (recommended)
df_list <- sheets_into_list(
excel_file = "data/my_experiment.xlsx",
clean_time = TRUE, # Convert time to Date format
remove_apnea = TRUE, # Remove apnea sheets
date_average = TRUE, # Average within each day
remove_na = TRUE # Remove rows with missing values
)Result: A list with 16 data frames (one per mouse)
# How many sheets loaded?
length(df_list)
# Sheet names
names(df_list)
# Look at first mouse
head(df_list[[1]])
# Check structure
str(df_list[[1]])
# Summary statistics
summary(df_list[[1]])Pro tip: Always inspect your data before analysis!
| Parameter | Full Name | What it Measures |
|---|---|---|
| f | Frequency | Breaths per minute |
| TVb | Tidal Volume | Volume per breath (mL) |
| MVb | Minute Ventilation | Total air per minute (mL/min) |
| Penh | Enhanced Pause | Airway resistance proxy |
| PAU | Pause | Breathing pause duration |
| Ti, Te | Inspiration/Expiration Time | Timing parameters |
| PIF, PEF | Peak Flows | Maximum flow rates |
# Method 1: Simple list
groups <- set_group_names(
"Uninfected WT",
"Uninfected Benac",
"Infected WT",
"Infected Benac"
)
# Method 2: From vector
group_names <- c("Uninfected WT", "Uninfected Benac",
"Infected WT", "Infected Benac")
groups <- set_group_names(group_names)
# View
print(groups)# Interactive assignment (best for first time)
mapping <- assign_groups_interactive(groups, df_list = df_list)You’ll see prompts like:
Enter sheet numbers for group 'Uninfected WT':
1-4
Enter sheet numbers for group 'Uninfected Benac':
5-8Tips: - Use ranges: 1-4 instead of
1,2,3,4 - Combine: 1-4,7,9-11 - Double-check
your assignments!
# If you already know the mapping
mapping <- list(
"Uninfected WT" = 1:4,
"Uninfected Benac" = 5:8,
"Infected WT" = 9:12,
"Infected Benac" = 13:16
)
# Verify
print(mapping)When to use: - You know your sheet order - Running analysis repeatedly - Scripting automated analysis
# Option 1: Keep nested structure (for individual mice)
organized <- organize_by_groups(
df_list = df_list,
group_mapping = mapping,
add_subject_id = TRUE
)
# Option 2: Combine within groups (for group analysis)
organized_combined <- organize_by_groups(
df_list = df_list,
group_mapping = mapping,
add_subject_id = TRUE,
combine_groups = TRUE
)
# Option 3: One big data frame (for PCA)
all_data <- organize_by_groups(
df_list = df_list,
group_mapping = mapping,
add_subject_id = TRUE,
combine_groups = TRUE,
combine_all = TRUE
)Nested (organized):
$`Uninfected WT`
$`Uninfected WT_1` # Mouse 1 data frame
$`Uninfected WT_2` # Mouse 2 data frame
...Combined (organized_combined):
$`Uninfected WT` # All 4 mice combined
$`Uninfected Benac` # All 4 mice combined
...All data (all_data):
One data frame with all 16 mice, labeled by group and subjectIndividual mice are noisy → Group averages show trends
For each time point: - Average all mice in each group - Get one value per group per parameter - Easier to visualize and compare
# Calculate means for each group at each time point
group_avgs <- calculate_group_averages(
df_list = df_list,
group_mapping = mapping,
time_var = "Time",
summary_fun = mean, # Could also use median
keep_n = TRUE # Include sample size
)
# Result: List with one data frame per group
names(group_avgs)
head(group_avgs[[1]])# Each group is a data frame
group_avgs$`Uninfected WT`
# Columns include:
# - Time: Date
# - f, TVb, MVb, etc.: Averaged parameters
# - n: Number of mice contributing (should be 4)
# - group: Group identifier
# Check sample sizes
unique(group_avgs$`Uninfected WT`$n) # Should be 4Red flag: If n varies a lot, you have
missing data issues!
# Basic time series plot
plots <- plot_wbp_timeseries(
data = group_avgs,
parameters = c("f", "TVb", "MVb", "Penh"),
plot_type = "by_parameter" # One plot per parameter
)
# View frequency plot
plots$f# With rolling average smoothing
plots <- plot_wbp_timeseries(
data = group_avgs,
parameters = c("f", "TVb", "Penh"),
smooth_method = "rolling",
smooth_window = 3, # 3-day window
line_size = 1.2,
show_points = FALSE
)
# LOESS smoothing (more sophisticated)
plots_loess <- plot_wbp_timeseries(
data = group_avgs,
parameters = c("f", "Penh"),
smooth_method = "loess",
smooth_span = 0.2, # Adjust smoothness
show_se = TRUE # Show confidence interval
)# Publication-ready time series
plots <- plot_wbp_timeseries(
data = group_avgs,
parameters = c("f", "TVb", "MVb", "Penh"),
plot_type = "by_parameter",
smooth_method = "rolling",
smooth_window = 3,
color_palette = "Set2", # Colorblind-friendly
line_size = 1.5,
show_points = FALSE,
save_plots = TRUE,
output_dir = "figures/timeseries",
file_prefix = "exp1",
dpi = 600 # High resolution
)# All parameters in one figure
combined <- plot_wbp_timeseries(
data = group_avgs,
plot_type = "combined", # Multi-panel
parameters = c("f", "TVb", "MVb", "Penh"),
smooth_method = "rolling",
smooth_window = 3,
facet_scales = "free_y", # Each parameter gets its own y-axis
save_plots = TRUE,
width = 12,
height = 8
)Pro tip: Combined plots are great for presentations!
Area Under the Curve summarizes the entire time series into one number
Why AUC? - Captures overall response magnitude - Easy to compare groups statistically - Standard approach in pharmacology/toxicology
When to use baseline correction: - Infection studies: measure change from day 0 - Treatment studies: compare pre vs post
# Basic AUC calculation
auc_results <- calculate_auc(
data = group_avgs,
time_var = "Time"
)
# View results
head(auc_results)Output:
group parameter auc
Uninfected WT f 27000.5
Uninfected WT TVb 8500.2
...# AUC normalized to control group (fold-change)
auc_results <- calculate_auc(
data = group_avgs,
normalize_to = "Uninfected WT", # Reference group
baseline_correct = FALSE
)
# Now includes fold_change column
head(auc_results)Output:
group parameter auc auc_normalized fold_change
Uninfected WT f 27000 1.00 1.00
Infected WT f 32400 1.20 1.20 # 20% increase!# Measure change from day 0 (infection studies)
auc_results <- calculate_auc(
data = group_avgs,
normalize_to = "Uninfected WT",
baseline_correct = TRUE, # Subtract day 0 values first
method = "trapezoid"
)When to use: - Infection studies (compare day 0 to day 7) - Treatment studies (before/after) - Any study where baseline varies between groups
When NOT to use: - Groups already matched at baseline - Steady-state measurements
# Basic bar plot
plots <- plot_auc_bars(
auc_results = auc_results,
parameters = c("f", "TVb", "Penh")
)
# View frequency plot
plots$f_auc# Publication-ready with statistics
plots <- plot_auc_bars(
auc_results = auc_results,
parameters = c("f", "TVb", "MVb", "Penh"),
group_order = c("Uninfected WT", "Uninfected Benac",
"Infected WT", "Infected Benac"),
value_type = "normalized", # Plot fold-change
y_axis_start = "smart", # Smart scaling
show_values = TRUE, # Show numbers on bars
show_stats = TRUE, # Add significance stars
reference_group = "Uninfected WT",
color_palette = "Set1",
save_plots = TRUE,
output_dir = "figures/auc",
file_prefix = "experiment1",
width = 6,
height = 4.5,
dpi = 600 # High resolution!
)Traditional (zero-based):
|
| 30000 ___
| 29000 | |
| 28000 | |
| ... | | (tiny differences hard to see)
| 1000 | |
| 0 |_|Smart scaling:
|
| 30000 ___
| 29500 | |
| 29000 | | (differences much clearer!)
| 28500 |_|When to use which: - Smart: When differences are small but meaningful - Zero: When journal requires it, or differences are huge
# Different color schemes
plots_viridis <- plot_auc_bars(
auc_results,
color_palette = "viridis" # Good for projectors
)
plots_bw <- plot_auc_bars(
auc_results,
color_palette = "grayscale" # Black & white journals
)
# Custom group order (logical arrangement)
plots_ordered <- plot_auc_bars(
auc_results,
group_order = c("Uninfected WT", "Infected WT",
"Uninfected Benac", "Infected Benac"),
# Groups WT vs Benac side-by-side
)# Automatic significance testing
plots <- plot_auc_bars(
auc_results,
show_stats = TRUE,
reference_group = "Uninfected WT", # Compare all to this
# Adds stars: * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001
)Note: These are placeholder p-values based on fold-change magnitude. For real statistics, you need: - Individual mouse data (not group means) - Proper t-tests or ANOVA - Multiple testing correction
For now: Stars show which groups differ substantially from control
# Basic heatmap
heatmap <- plot_auc_heatmap(
auc_results = auc_results,
value_type = "normalized", # Plot fold-change
cluster_rows = TRUE, # Cluster parameters
cluster_cols = TRUE # Cluster groups
)What it shows: - Rows = Parameters - Columns = Groups - Colors = Fold-change vs control - Clustering reveals patterns
# Publication-ready heatmap
heatmap <- plot_auc_heatmap(
auc_results = auc_results,
value_type = "normalized",
include_parameters = c("f", "TVb", "MVb", "Penh",
"PAU", "Ti", "Te"), # Select key params
cluster_rows = TRUE,
cluster_cols = TRUE,
clustering_method = "ward.D2", # Better clustering
color_scheme = "RdBu", # Red-Blue diverging
show_values = TRUE, # Show fold-change in cells
cell_width = 50,
cell_height = 40,
font_size = 12,
save_plot = TRUE,
output_dir = "figures/heatmaps",
dpi = 300
)Colors: - Red = Increased vs control (>1.0 fold) - White = No change (1.0 fold) - Blue = Decreased vs control (<1.0 fold)
Clustering: - Parameters that cluster together respond similarly - Groups that cluster together have similar profiles
Example insights: - “f, TVb, MVb cluster together” → Breathing effort parameters linked - “Infected groups cluster away from Uninfected” → Clear infection effect
# Remove parameters that don't change much
heatmap <- plot_auc_heatmap(
auc_results,
exclude_parameters = c("Comp", "EEP", "EIP"), # Low info
# Or use include_parameters for positive selection
)
# Only show respiratory effort parameters
heatmap_effort <- plot_auc_heatmap(
auc_results,
include_parameters = c("f", "TVb", "MVb", "Penh", "PAU")
)Pro tip: Focus heatmap on your key parameters for clarity!
Principal Component Analysis reduces many parameters to 2-3 dimensions
Why PCA? - See overall patterns - Identify outliers - Check if groups naturally separate - Explore which parameters drive differences
Each point = one mouse Distance between points = similarity in respiratory profile
# Basic PCA (colored by experimental group)
pca_result <- plot_pca(
data = df_list, # Use individual mice
group_mapping = mapping,
show_loadings = FALSE,
save_plot = FALSE
)
# View the plot
pca_result$plot
# Check variance explained
print(pca_result$variance)Typical output:
PC1 explains 48.6% of variance
PC2 explains 29.4% of variance
Total: 78.0%# Show which parameters drive the separation
pca_result <- plot_pca(
data = df_list,
group_mapping = mapping,
show_loadings = TRUE, # Show parameter vectors
n_loadings = 5, # Top 5 contributors
show_ellipses = TRUE, # 95% confidence ellipses
color_palette = "Set2"
)Interpretation: - Arrows point in direction of increasing parameter value - Long arrows = parameter contributes more to separation - Example: If “Penh” arrow points toward Infected groups → Penh is elevated in infected mice
# See if unsupervised clustering matches experimental groups
pca_result <- plot_pca(
data = df_list,
group_mapping = mapping,
use_clustering = TRUE,
num_clusters = 4, # Try 4 clusters
color_by = "cluster", # Color by clusters
show_ellipses = FALSE # Turn off (not enough per cluster)
)
# Compare clusters to experimental groups
table(Experimental = pca_result$scores$group,
Cluster = pca_result$scores$cluster)Good clustering: K-means matches your experimental groups → groups are truly distinct!
# Plot PC3 vs PC4 (not just PC1 vs PC2)
pca_34 <- plot_pca(
data = df_list,
group_mapping = mapping,
components = c("PC3", "PC4"),
show_loadings = TRUE
)
# Check what they explain
pca_result$varianceWhen to look at PC3/PC4: - PC1/PC2 don’t separate groups well - Looking for secondary effects - Checking for batch effects
Tight clusters = Group is homogeneous (good!)
Spread out points = High within-group variability (concerning)
Groups overlap = No clear difference between groups
Groups separate = Clear biological difference
Outliers = Check that mouse: - Technical problem? - Biological responder? - Wrong group assignment?
# ===== SETUP =====
library(plethR)
# ===== IMPORT DATA =====
df_list <- sheets_into_list(
"data/experiment.xlsx",
clean_time = TRUE,
date_average = TRUE,
remove_apnea = TRUE
)
# ===== ORGANIZE GROUPS =====
groups <- set_group_names("Uninfected WT", "Uninfected Benac",
"Infected WT", "Infected Benac")
mapping <- assign_groups_interactive(groups, df_list)
# ===== CALCULATE AVERAGES =====
group_avgs <- calculate_group_averages(df_list, mapping)
# (continued on next slide)# ===== TIME SERIES PLOTS =====
ts_plots <- plot_wbp_timeseries(
group_avgs,
parameters = c("f", "TVb", "Penh"),
smooth_method = "rolling",
smooth_window = 3,
save_plots = TRUE,
dpi = 600
)
# ===== AUC ANALYSIS =====
auc <- calculate_auc(group_avgs, normalize_to = "Uninfected WT")
auc_plots <- plot_auc_bars(
auc,
group_order = groups,
show_stats = TRUE,
reference_group = "Uninfected WT",
save_plots = TRUE,
dpi = 600
)
# (continued on next slide)# ===== HEATMAP =====
heatmap <- plot_auc_heatmap(
auc,
include_parameters = c("f", "TVb", "MVb", "Penh"),
color_scheme = "RdBu",
save_plot = TRUE
)
# ===== PCA =====
pca <- plot_pca(
df_list,
group_mapping = mapping,
show_loadings = TRUE,
save_plot = TRUE
)
# ===== EXPORT =====
export_to_excel(group_avgs, "group_averages", "results")
export_to_excel(auc, "auc_results", "results")
print("Analysis complete!")Save your workflow as an R script:
# analysis_YYYYMMDD.R
# Purpose: Analyze experiment ABC
# Date: 2024-11-19
# Author: Your Name
# Load packages
library(plethR)
library(dplyr) # For data manipulation
# Set parameters
EXPERIMENT <- "exp1"
DATA_FILE <- "data/exp1_data.xlsx"
CONTROL_GROUP <- "Uninfected WT"
# ... rest of analysis ...Benefits: - Reproducible - Easy to re-run - Can share with colleagues - Document your decisions
Before you start: 1. Check one Excel file per experiment 2. Check consistent sheet naming 3. Check clean date/time formats 4. Check metadata noted elsewhere
Folder structure:
Project/
├── data/
│ └── experiment.xlsx
├── figures/
│ ├── timeseries/
│ ├── auc/
│ └── pca/
├── results/
└── analysis.RFor figures: - [ ] Use 600 DPI for journals - [ ] Consistent color schemes across figures - [ ] Proper group ordering (control first) - [ ] Appropriate y-axis scaling - [ ] Statistical annotations where needed - [ ] Clear axis labels and titles
For data: - [ ] Export processed data - [ ] Document any exclusions - [ ] Save group assignments - [ ] Keep raw data separate
Not checking group assignments → always verify with
print(mapping)
*Forgetting to clean time** → use clean_time = TRUE
*Using wrong reference group** → double-check
normalize_to matches your control
*Not averaging replicates** → se date_average = TRUE for
daily data
*Mixing infected and uninfected baseline data** → onsider
baseline_correct = TRUE
“Function not found”
“No numeric columns found”
“Groups have < 3 subjects, skipping ellipses”
# This is just a warning - plots still work
# Turn off ellipses if needed
plot_auc_bars(..., show_ellipses = FALSE)In R:
# Package overview
?plethR
# Specific function
?calculate_auc
# Browse all help files
help(package = "plethR")
# Vignette
vignette("plethR-workflow")Online: - GitHub: https://github.com/camronp/plethR - Issues: https://github.com/camronp/plethR/issues - Email: [camronpearce@gmail.com]
# Use your lab's colors
my_colors <- c("#FF6B6B", "#4ECDC4", "#45B7D1", "#FFA07A")
# Currently plethR doesn't support custom colors directly
# But you can modify plots after creation
# Example with ggplot2:
library(ggplot2)
# Get a plot
p <- auc_plots$f_auc
# Modify colors
p + scale_fill_manual(values = my_colors) +
scale_color_manual(values = my_colors)# Process multiple experiments at once
experiments <- c("exp1", "exp2", "exp3")
for (exp in experiments) {
# Import
df <- sheets_into_list(paste0("data/", exp, ".xlsx"),
clean_time = TRUE,
date_average = TRUE)
# Analyze
mapping <- list(
"Control" = 1:4,
"Treatment" = 5:8
)
avgs <- calculate_group_averages(df, mapping)
auc <- calculate_auc(avgs, normalize_to = "Control")
# Save
plot_auc_bars(auc,
save_plots = TRUE,
output_dir = paste0("figures/", exp),
file_prefix = exp)
export_to_excel(auc, paste0(exp, "_results"), "results")
}# Focus on respiratory effort parameters only
effort_params <- c("f", "TVb", "MVb")
# Use throughout analysis
ts_plots <- plot_wbp_timeseries(
group_avgs,
parameters = effort_params
)
auc_plots <- plot_auc_bars(
auc_results,
parameters = effort_params
)
heatmap <- plot_auc_heatmap(
auc_results,
include_parameters = effort_params
)# Compare specific time points (e.g., Day 0 vs Day 7)
day0 <- group_avgs %>%
lapply(function(df) df %>% filter(Time == min(Time)))
day7 <- group_avgs %>%
lapply(function(df) df %>% filter(Time == max(Time)))
# Calculate fold-change for each parameter
changes <- mapply(function(d0, d7) {
data.frame(
parameter = names(d0)[-1],
day0 = as.numeric(d0[-1]),
day7 = as.numeric(d7[-1]),
fold_change = as.numeric(d7[-1]) / as.numeric(d0[-1])
)
}, day0, day7, SIMPLIFY = FALSE)# Analyze only specific time window
library(dplyr)
# Days 3-7 only (post-infection)
group_avgs_filtered <- lapply(group_avgs, function(df) {
df %>%
filter(Time >= as.Date("2024-01-03") &
Time <= as.Date("2024-01-07"))
})
# Recalculate AUC for this window
auc_postinfection <- calculate_auc(
group_avgs_filtered,
normalize_to = "Uninfected WT"
)# plethR provides visualization, but for proper statistics:
# You need individual mouse data (not group means)
# Example: Compare AUC between two groups
library(dplyr)
# Get individual AUC for each mouse (you'd calculate this from df_list)
# Placeholder example:
auc_individual <- data.frame(
mouse_id = 1:16,
group = rep(c("Control", "Treated"), each = 8),
auc_f = rnorm(16, mean = 27000, sd = 2000)
)
# T-test
t.test(auc_f ~ group, data = auc_individual)
# ANOVA for multiple groups
anova_result <- aov(auc_f ~ group, data = auc_individual)
summary(anova_result)
# Post-hoc tests
TukeyHSD(anova_result)Note: For publication, always use proper statistical tests on individual data!
# Create a summary table for your paper
library(dplyr)
summary_table <- auc_results %>%
select(group, parameter, auc, auc_normalized) %>%
filter(parameter %in% c("f", "TVb", "Penh")) %>%
arrange(parameter, group)
# Export as Excel
export_to_excel(summary_table, "table1_auc_summary", "results")
# Or format for Word/LaTeX
library(knitr)
kable(summary_table, digits = 2, caption = "AUC Results by Group")For statistics:
library(emmeans) # Estimated marginal means
library(multcomp) # Multiple comparisons
library(lme4) # Mixed modelsFor advanced visualization:
library(patchwork) # Combine multiple plots
library(cowplot) # Publication-ready plot arrangements
library(gganimate) # Animated time seriesFor reports:
Question: Does infection affect breathing frequency?
# Import data
df_list <- sheets_into_list("infection_study.xlsx",
clean_time = TRUE,
date_average = TRUE)
# Groups: Uninfected vs Infected (4 mice each)
groups <- set_group_names("Uninfected", "Infected")
mapping <- list("Uninfected" = 1:4, "Infected" = 5:8)
# Calculate averages
avgs <- calculate_group_averages(df_list, mapping)
# Plot time series - see when frequency changes
plot_wbp_timeseries(
avgs,
parameters = "f",
smooth_method = "loess",
save_plots = TRUE
)
# Calculate AUC with baseline correction
auc <- calculate_auc(avgs,
baseline_correct = TRUE,
normalize_to = "Uninfected")
# Visualize
plot_auc_bars(auc,
parameters = "f",
show_stats = TRUE,
reference_group = "Uninfected")Result: Can quantify the infection-induced change in breathing!
Question: Does drug dose affect airway resistance (Penh)?
# Groups: Vehicle, Low, Medium, High dose
groups <- set_group_names("Vehicle", "0.1mg", "1mg", "10mg")
mapping <- list(
"Vehicle" = 1:4,
"0.1mg" = 5:8,
"1mg" = 9:12,
"10mg" = 13:16
)
avgs <- calculate_group_averages(df_list, mapping)
# AUC analysis
auc <- calculate_auc(avgs, normalize_to = "Vehicle")
# Plot with logical ordering
plot_auc_bars(
auc,
parameters = "Penh",
group_order = c("Vehicle", "0.1mg", "1mg", "10mg"),
show_stats = TRUE,
reference_group = "Vehicle"
)
# Check for dose-response relationship
auc %>%
filter(parameter == "Penh") %>%
arrange(group)Question: Does treatment work differently in WT vs KO mice?
# 2x2 design: WT/KO × Vehicle/Drug
groups <- set_group_names(
"WT Vehicle", "WT Drug",
"KO Vehicle", "KO Drug"
)
mapping <- list(
"WT Vehicle" = 1:4,
"WT Drug" = 5:8,
"KO Vehicle" = 9:12,
"KO Drug" = 13:16
)
avgs <- calculate_group_averages(df_list, mapping)
auc <- calculate_auc(avgs, normalize_to = "WT Vehicle")
# Organize for comparison
plot_auc_bars(
auc,
group_order = c("WT Vehicle", "WT Drug",
"KO Vehicle", "KO Drug"),
parameters = c("f", "Penh"),
value_type = "normalized"
)
# PCA to see overall patterns
plot_pca(df_list, group_mapping = mapping, show_loadings = TRUE)Question: Do mice recover after treatment ends?
# Groups measured at: Baseline, Treatment, Recovery
# Import data with all time points
df_list <- sheets_into_list("recovery_study.xlsx",
clean_time = TRUE)
# Define time windows
baseline <- df_list %>%
lapply(function(df) filter(df, Time <= as.Date("2024-01-07")))
treatment <- df_list %>%
lapply(function(df) filter(df, Time > as.Date("2024-01-07") &
Time <= as.Date("2024-01-14")))
recovery <- df_list %>%
lapply(function(df) filter(df, Time > as.Date("2024-01-14")))
# Calculate AUC for each phase
auc_baseline <- calculate_auc(baseline, mapping)
auc_treatment <- calculate_auc(treatment, mapping)
auc_recovery <- calculate_auc(recovery, mapping)
# Compare phases
combined_auc <- bind_rows(
auc_baseline %>% mutate(phase = "Baseline"),
auc_treatment %>% mutate(phase = "Treatment"),
auc_recovery %>% mutate(phase = "Recovery")
)Q: How many mice do I need per group? A: Minimum 3 for statistics, 4-6 is standard, more is better for power
Q: What if I have missing time points? A: plethR
handles this it averages what’s available. Use
remove_na = TRUE
Q: Can I analyze human data? A: No, this is for mouse wbp data only
Q: How do I cite plethR? A: See GitHub README for citation format
Q: What if my data format is different? A: Contact me we can help adapt the package and push updates
Time Series: - See how parameters change over time - Identify when effects occur - Compare group trajectories
AUC Bar Plots: - Statistical comparisons between groups - Summarize overall response - Publication main figures
Heatmaps: - Compare many parameters at once - Discover unexpected patterns - Supplementary figures
PCA: - Overall group separation - Identify outliers - Exploratory analysis
Practice: 1. Load your own data 2. Run through the workflow 3. Experiment with different parameters 4. Save your script
Share: - Teach others in the lab - Contribute back to plethR!
Learn More: - Read function documentation:
?function_name - Check the vignette:
vignette("plethR-workflow") - Visit GitHub: https://github.com/camronp/plethR
plethR: - GitHub: https://github.com/camronp/plethR - Installation:
devtools::install_github("camronp/plethR") -
Issues/Questions: Open a GitHub issue
R Resources: - R for Data Science: https://r4ds.had.co.nz/ - ggplot2 book: https://ggplot2-book.org/ - Stack Overflow: Tag your questions with [r] and [plethR]
Statistical Help: - Consult with CSU your statistics core - Consider a biostatistician for more complex study designs
I want to hear from you!
What works well? What’s confusing? What features are missing?
How to give feedback: 1. GitHub Issues: https://github.com/camronp/plethR/issues 2. Email: camronpearce@gmail.com 3. In person: Stop by the lab!
Contributing: - Find a bug or want to feature an idea? Let me know!
Package Developer: Cam
Email: camronpearce@gmail.com GitHub: https://github.com/camronp/plethR
On GitHub: - Complete documentation - Example datasets in vignette
In the Lab: - Come find me in the next two weeks
Questions?
# Import
sheets_into_list(file, clean_time = TRUE, date_average = TRUE)
# Organize
set_group_names("Group1", "Group2", ...)
assign_groups_interactive(groups, df_list)
organize_by_groups(df_list, mapping, combine_groups = TRUE)
# Analyze
calculate_group_averages(df_list, mapping)
calculate_auc(data, normalize_to = "Control", baseline_correct = TRUE)
# Visualize
plot_wbp_timeseries(data, parameters, smooth_method = "rolling")
plot_auc_bars(auc_results, show_stats = TRUE, reference_group = "Control")
plot_auc_heatmap(auc_results, color_scheme = "RdBu")
plot_pca(df_list, group_mapping, show_loadings = TRUE)
# Export
export_to_excel(data, "filename", "output_dir")Ctrl/Cmd + Enter: Run current lineCtrl/Cmd + Shift + Enter: Run entire chunkCtrl/Cmd + Shift + C: Comment/uncommentCtrl/Cmd + Shift + M: Insert pipe
%>%Tab: Auto-completeCtrl/Cmd + Shift + F10: Restart RColorBrewer: - "Set1": Bold, distinct
colors (default) - "Set2": Softer, pastel colors -
"Dark2": Dark, saturated colors - "Paired":
Pairs of related colors
Viridis: - "viridis": Blue to yellow
(default) - "plasma": Purple to yellow -
"magma": Black to white - "inferno": Black to
yellow
For publications: - "Set2":
Colorblind-friendly, professional - "grayscale": Black
& white journals
Project_Name/
├── data/
│ ├── raw/
│ │ └── experiment.xlsx
│ └── processed/
├── scripts/
│ ├── 01_import.R
│ ├── 02_analysis.R
│ └── 03_figures.R
├── figures/
│ ├── timeseries/
│ ├── auc/
│ ├── heatmaps/
│ └── pca/
├── results/
│ └── tables/
├── docs/
│ └── lab_notebook.md
└── Project_Name.Rproj| Code | Parameter | Units | Meaning |
|---|---|---|---|
| f | Frequency | breaths/min | Breathing rate |
| TVb | Tidal Volume | mL | Air per breath |
| MVb | Minute Ventilation | mL/min | Total air/min |
| Penh | Enhanced Pause | unitless | Airway resistance |
| PAU | Pause | ms | Breathing pause |
| Ti | Inspiratory Time | s | Inhalation duration |
| Te | Expiratory Time | s | Exhalation duration |
| PIF | Peak Inspiratory Flow | mL/s | Max inflow |
| PEF | Peak Expiratory Flow | mL/s | Max outflow |
Thank you for your attention!
Remember: 1. Practice with your own data 2. Save your scripts 3. Ask questions 4. Share with colleagues 5. Provide feedback
See you at the next workshop!