Visualising results
In greatR: Gene Registration from Expression and Time-Courses in R

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

After running registration function scale_and_register_data() as shown in the registering data article, users can summarise and visualise the results as illustrated in the figure below.

knitr::include_graphics("figures/visualisation_process.png")

Get summary from registration results

The total number of registered and non-registered genes can be obtained by running function summary_model_comparison() with registration_results$model_comparison_df as an input.

# Load the package
library(greatR)
library(dplyr)

all_data_df <- system.file("extdata/brapa_arabidopsis_all_replicates.csv", package = "greatR") %>%
  utils::read.csv()

# Running the registration
registration_results <- scale_and_register_data(
  input_df = all_data_df,
  stretches = c(3, 2.5, 2, 1.5, 1),
  shifts = seq(-4, 4, length.out = 33),
  min_num_overlapping_points = 4,
  initial_rescale = FALSE,
  do_rescale = TRUE,
  accession_data_to_transform = "Col0",
  accession_data_ref = "Ro18",
  start_timepoint = "reference"
)

Function summary_model_comparison() returns a list which contains three different objects:

df_summary contains result summaries of the registration results,
registered_genes is list of gene accessions which were successfully registered, and
non_registered_genes is a list of non-registered gene accessions.

# Get all of summary
all_summary <- summary_model_comparison(registration_results$model_comparison_df)

all_summary$df_summary %>%
  knitr::kable()

The list of gene accessions which were registered can be viewed by calling:

all_summary$registered_genes

Plot registration results

Function plot_registration_results() allows users to plot registration results of the genes of interest.

# Plot registration result
plot_registration_results(
  registration_results$imputed_mean_df,
  ncol = 3
)

Users also have an option to include information or label on the plot whether particular genes are registered or not, as well as the registration parameters by include model comparison data frame as shown below.

# Plot registration result
plot_registration_results(
  registration_results$imputed_mean_df,
  registration_results$model_comparison_df,
  ncol = 3,
  sync_timepoints = TRUE
)

Notice that to only include same time points between samples, users can set sync_timepoints = TRUE.

Analyse similarity of expression profiles overtime before and after registering

Calculate sample distance

After registering the sample data, users can compare the overall similarity before and after registering using the function calculate_between_sample_distance().

sample_distance <- calculate_between_sample_distance(
  registration_results,
  accession_data_ref = "Ro18"
)

Function calculate_between_sample_distance() returns a list of seven data frames:

distance_mean_df is distance of mean expression values.
distance_scaled_mean_df is distance of scaled mean expression (all genes).
distance_scaled_mean_df_only_nonreg is distance of scaled mean expression (only not-registered genes).
distance_scaled_mean_df_only_reg is distance of scaled mean expression (only registered genes).
distance_registered_df is distance of registered & scaled mean expression (all genes).
distance_registered_df_only_reg is distance of registered & scaled mean expression (only registered genes).

Plot heatmap of sample distances

Each of these data frames above can be visualised using function plot_heatmap().

# Plot heatmap of mean expression profiles distance before scaling
plot_heatmap(sample_distance$distance_mean_df)

# Plot heatmap of mean expression profiles distance after scaling
plot_heatmap(sample_distance$distance_scaled_mean_df)

# Plot heatmap of mean expression profiles distance after registration process
plot_heatmap(
  sample_distance$distance_registered_df_only_reg, 
  same_max_timepoint = TRUE, 
  same_min_timepoint = TRUE
)