demonanalysis

Analysis and plotting of demon data

From a single simulation

To plot charts of variant allele frequencies and genotype sizes use plot_all_charts. For normal tissue, you should specify a value for max_allele_count so that the axes are appropriately scaled.

To plot Muller plots and grids use plot_all_images.

From a batch of simulations run on a cluster

Updating

To pull the latest github version into a cluster folder (using terminal):

cd XXX # replace XXX with the path to the cluster folder
git pull https://github.com/robjohnnoble/demonanalysis

To install that version on a cluster profile (using R):

library(devtools)
install("XXX/demonanalysis") # replace XXX with the path to the cluster folder

Set up

library(demonanalysis)

subfolder_name <- "XXX" # insert name of top-level folder

input_dir <- paste0("all_results/", subfolder_name) # folder containing results of the batch
num_parameters <- count_parameters(input_dir) # number of simulation parameters (first columns in data)
output_dir_plots <- paste0("plots/", subfolder_name) # folder to receive image files
output_dir_data <- paste0("data/", subfolder_name) # folder containing data files

Check which simulations completed successfully

all_statuses(input_dir, summary = TRUE) # should be "Exit code 0" (when finished) or "So far no status" (while running)

Create plots for a batch of simulations

For growing tumours, to plot charts of variant allele frequencies and genotype sizes:

create_plots_batch(input_dir, output_dir = output_dir_plots, type = "chart")

For normal tissue, to plot charts of variant allele frequencies and genotype sizes:

create_plots_batch(input_dir, output_dir = output_dir_plots, type = "chart", max_size = 50)

To plot Muller plots and grids:

create_plots_batch(input_dir, output_dir = output_dir_plots, type = "plot")

Get general-purpose data

data <- all_output(input_dir, include_diversities = FALSE) # combined data for a batch of simulations, excluding diversity columns

Plot relationships between variables (using dataframe)

plot_curves_faceted(data, num_parameters, x_var = "Generation", y_var = "MeanBirthRate", output_filename = "curves", output_dir = output_dir_plots)
# change x_var, y_var and output_filename as appropriate

Get additional data for forecasting

data <- all_output(input_dir) # combined data for a batch of simulations, including diversity columns
data <- add_relative_time(data, start_size = 5500, num_parameters = num_parameters) # add columns useful for plotting trajectories

start_size_range <- 500 * 2^(0:8) # NumCells at time of initial measurement for forecasting
gap_range <- (1:10)/10 # gap between time of initial measurement and second measurement
final_size <- 1E5 # waiting time is measured until tumour reaches this NumCells value

summary <- get_summary(data, start_size_range, gap_range, final_size, num_parameters = num_parameters) # summary data for each simulation, for each combination of gap and final_size

cor_summary <- get_cor_summary(summary, c("DriverDiversity", "DriverEdgeDiversity"), num_parameters = num_parameters, min_count = 5) # summary dataframe of correlations with "outcome", including all cells
wait_cor_summary <- get_wait_cor_summary(summary, c("DriverDiversity", "DriverEdgeDiversity"), num_parameters = num_parameters, min_count = 5) # summary dataframe of correlations with "waiting_time", including all cells
depth_wait_cor_summary <- get_wait_cor_summary(summary, c(paste0("DriverDiversityFrom1SamplesAtDepth", 0:10), paste0("DriverDiversityFrom4SamplesAtDepth", 0:10)), num_parameters, min_count = 5) # summary dataframe of correlations with "waiting_time" for different biopsy protocols