Beautiful plots made via myTAI

Gallery of different plots available with myTAI

Here are example outputs of plotting functions from myTAIv2.

We hope these plots can inspire your analysis!

Bulk RNA-seq data

example_phyex_set is an example BulkPhyloExpressionSet object.

To learn more about bringing your dataset into myTAI, follow this vignette here:
→ 📊

library(myTAI); library(S7); library(ggplot2); library(patchwork)
data("example_phyex_set")

myTAI plots can be modified as a ggplot2 object.

myTAI::plot_signature(example_phyex_set, 
                      show_p_val = TRUE, 
                      conservation_test = stat_flatline_test,
                      colour = "lavender") +
  # as the plots are ggplot2 objects, we can simply modify them using ggplot2
  ggplot2::labs(title = "Developmental stages of A. thaliana")

plot_signature function output with stat_flatline_test

module_info <- list(early = 1:3, mid = 4:6, late = 7:8)
myTAI::plot_signature(example_phyex_set,
                      show_p_val = TRUE,
                      conservation_test = stat_reductive_hourglass_test,
                      modules = module_info,
                      colour = "lavender")

plot_signature function output with stat_reductive_hourglass_test

Transformation and robustness checks

See more here:
→ 🛡️

myTAI::plot_signature_transformed(
  example_phyex_set)

## Computing: [========================================] 100% (done)                         
## Computing: [========================================] 100% (done)                         
## Computing: [========================================] 100% (done)                         
## Computing: [========================================] 100% (done)                         
## Computing: [========================================] 100% (done)

plot_signature_transformed function output

myTAI::plot_signature_gene_quantiles(
  example_phyex_set)

plot_signature_gene_quantiles function output

Statistical tests and plotting results

See more here:
→ 📈

myTAI::stat_flatline_test(
  example_phyex_set, plot_result = TRUE)

stat_flatline_test function output

## 
## Statistical Test Result
## =======================
## Method: Flat Line Test 
## Test statistic: 0.0446168 
## P-value: 0.4367048 
## Alternative hypothesis: greater 
## Data: Embryogenesis 2019

res_flt <- myTAI::stat_flatline_test(example_phyex_set, plot_result = FALSE)

myTAI::plot_cullen_frey(res_flt)

plot_cullen_frey function output for stat_flatline_test

## summary statistics
## ------
## min:  0.0008681294   max:  0.9345385 
## median:  0.04342944 
## mean:  0.08851276 
## estimated sd:  0.160333 
## estimated skewness:  3.981776 
## estimated kurtosis:  20.7357

myTAI::plot_null_txi_sample(res_flt) +
  ggplot2::guides(x =  guide_axis(angle = 90))

plot_null_txi_sample function output for stat_flatline_test

module_info <- list(early = 1:3, mid = 4:6, late = 7:8)
myTAI::stat_reductive_hourglass_test(
  example_phyex_set, plot_result = TRUE,
  modules = module_info)

stat_reductive_hourglass_test function output

## 
## Statistical Test Result
## =======================
## Method: Reductive Hourglass Test 
## Test statistic: -0.03191036 
## P-value: 0.1736998 
## Alternative hypothesis: greater 
## Data: Embryogenesis 2019

Average gene expression level by phylostratum

myTAI::plot_strata_expression(example_phyex_set)

plot_strata_expression function output

plot_strata_expression with scaled y axis

myTAI::plot_strata_expression(example_phyex_set) +
  ggplot2::scale_y_log10() +
  ggplot2::labs(x = "Expression aggregated by mean (log-scaled)")

plot_strata_expression function output ggplot2

plot_strata_expression with explicit transformation

library(patchwork)
p1 <- myTAI::plot_strata_expression(example_phyex_set |> myTAI::tf(log1p))

# equivalent to 
p2 <- example_phyex_set |> myTAI::tf(log1p) |> myTAI::plot_strata_expression() 

p1+p2

plot_strata_expression function output

As you can see, both plots are identical. This example demonstrates that there are multiple ways to achieve the same result through piping (|>) operator in R. |> is basically the same as %>%.

Contribution to the overall TAI by phylostratum

myTAI::plot_contribution(example_phyex_set)

plot_contribution function output

Curious about methods to obtain gene age information? See more here:
→ 📚

For other analogous methods to assign evolutionary or expression information to each gene for TDI, TSI etc., see here:
→ 🧬

myTAI::plot_distribution_expression(example_phyex_set)

plot_distribution_expression function output

Contribution to the overall TAI by partial TAI (pTAI)

pTAI, or

\[ \mathrm{pTAI}_i = \frac{\mathrm{ps}_i \cdot e_{is}}{\sum_{i=1}^{n} e_{is}} \]

where \(e_{is}\) denotes the expression level of a given gene \() i\) in sample \(s\), \({ps}_i\) is its gene age assignment, and \(n\) is the total number of genes, is the per-gene contribution to the overall TAI. (Summing pTAI across all genes gives in a given sample \(s\) gives the overall \({TAI}_s\) )

pTAI QQ plot compares the partial TAI distributions of various developmental stages against a reference stage (default is stage 1).

myTAI::plot_distribution_pTAI_qqplot(example_phyex_set)

plot_distribution_pTAI_qqplot function output

Phylostratum distribution

myTAI::plot_distribution_strata(example_phyex_set@strata) /
myTAI::plot_distribution_strata(
  example_phyex_set@strata,
  selected_gene_ids = myTAI::genes_top_variance(example_phyex_set, top_p = 0.95),
  as_log_obs_exp = TRUE
) + plot_annotation(title = "Distribution of gene ages (top), Observed vs Expected plot of top 5% variance genes (bottom)")

plot_distribution_strata function output

Expression heatmap

myTAI::plot_gene_heatmap(example_phyex_set)

plot_gene_heatmap function output default

myTAI::plot_gene_heatmap(example_phyex_set, cluster_rows = TRUE, show_reps=TRUE, show_gene_ids=TRUE, top_p=0.005)

plot_gene_heatmap function output clustered

myTAI::plot_gene_heatmap(example_phyex_set, cluster_rows = TRUE, show_reps=TRUE, top_p=0.005, std=FALSE, show_gene_ids=TRUE)

plot_gene_heatmap function output nonstd

Dimension reduction

At the gene level

myTAI::plot_gene_space(example_phyex_set)

plot_gene_space function output

myTAI::plot_gene_space(example_phyex_set,colour_by = "strata")

plot_gene_space function output by strata

At the sample level

myTAI::plot_sample_space(example_phyex_set) | myTAI::plot_sample_space(example_phyex_set, colour_by = "TXI")

plot_sample_space function output by TXI

# we can even do a UMAP
myTAI::plot_sample_space(example_phyex_set, method = "UMAP")

plot_sample_space function output by TXI

Inspecting mean-variance relationship

# highlighting top variance genes
top_var_genes <- myTAI::genes_top_variance(example_phyex_set, top_p = 0.9995)
p1 <- myTAI::plot_mean_var(example_phyex_set)
p2 <- myTAI::plot_mean_var(example_phyex_set, 
                     highlight_genes = top_var_genes)

p1 + p2 + plot_annotation(title = "Mean-variance: simple vs. highlighted top variance genes")

plot_mean_var function output simple vs highlighted

# with log transform and colouring by phylostratum
myTAI::plot_mean_var(example_phyex_set |> myTAI::tf(log1p), 
                     colour_by = "strata") +
  ggplot2::guides(colour = guide_legend(ncol=2))

plot_gene_space function output log transform coloured by strata

Individual gene expression profiles

# side by side: manual coloring vs strata coloring
p1 <- myTAI::plot_gene_profiles(example_phyex_set, max_genes = 10, colour_by = "manual")
p2 <- myTAI::plot_gene_profiles(example_phyex_set, max_genes = 10, colour_by = "strata")

p1 + p2 + plot_annotation(title = "Gene profiles: manual vs. strata coloring")

plot_gene_profiles function output manual vs strata coloring

# stage colouring with standardized log transformation
myTAI::plot_gene_profiles(example_phyex_set, max_genes = 10, 
                          transformation = "std_log", colour_by = "stage")

plot_gene_profiles function output stage std_log transformation

# faceted by phylostratum
myTAI::plot_gene_profiles(example_phyex_set, max_genes = 1000, 
                          colour_by = "strata", facet_by_strata = TRUE, show_set_mean = TRUE,
                          show_labels = FALSE)

plot_gene_profiles function output faceted by strata

These plots are examples of plots that myTAIv2 can generate. To check out the functions, use ? before the function (i.e. ?myTAI::plot_mean_var().

You can also find a list of plotting functions in Reference.

Single cell RNA-seq data

Most of the plotting functions shown above also apply for single cell RNA-seq data, as long as it is a ScPhyloExpressionSet object.

Let’s create an example single-cell dataset and explore the plotting capabilities:

# Load example single-cell data
data(example_phyex_set_sc)

example_phyex_set_sc

## PhyloExpressionSet object
## Class: myTAI::ScPhyloExpressionSet 
## Name: Single Cell Example 
## Species: Example Species 
## Index type: TXI 
## Cell Type : TypeA, TypeB, TypeC 
## Number of genes: 99 
## Number of cell type : 3 
## Number of phylostrata: 10 
## Total cells: 200 
## Cells per type:
## 
## TypeA TypeB TypeC 
##    70    66    64 
## Available metadata:
##   groups: TypeA, TypeB, TypeC
##   day: Day1, Day3, Day5, Day7
##   condition: Control, Treatment
##   batch: Batch1, Batch2, Batch3

# Check available identities
cat("Available identities for plotting:\n")

## Available identities for plotting:

print(example_phyex_set_sc@available_idents)

## [1] "groups"    "day"       "condition" "batch"

# Set up custom color schemes for better visualization
day_colors <- c("Day1" = "#3498db", "Day3" = "#2980b9", "Day5" = "#1f4e79", "Day7" = "#0d2a42")
condition_colors <- c("Control" = "#27ae60", "Treatment" = "#e74c3c")
group_colors <- c("TypeA" = "#e74c3c", "TypeB" = "#f39c12", "TypeC" = "#9b59b6")

example_phyex_set_sc@idents_colours[["day"]] <- day_colors
example_phyex_set_sc@idents_colours[["condition"]] <- condition_colors
example_phyex_set_sc@idents_colours[["groups"]] <- group_colors

Single-cell signature plots

# Basic signature plot showing TXI distribution across cell types
myTAI::plot_signature(example_phyex_set_sc)

plot_signature single-cell basic

# Plot without showing individual cells (just means)
myTAI::plot_signature(example_phyex_set_sc, show_reps = FALSE)

plot_signature single-cell without individual cells

# Plot TXI distribution by developmental day instead of cell type
myTAI::plot_signature(example_phyex_set_sc, primary_identity = "day", show_p_val = FALSE)

plot_signature single-cell by day

# Plot TXI distribution by experimental condition
myTAI::plot_signature(example_phyex_set_sc, primary_identity = "condition", show_p_val=FALSE)

plot_signature single-cell by condition

You can use a secondary identity for either coloring or faceting to create more informative plots:

# Plot by day, colored by condition
myTAI::plot_signature(example_phyex_set_sc, 
                     primary_identity = "day", 
                     secondary_identity = "condition",
                     show_p_val=FALSE)

plot_signature single-cell with secondary coloring

# Plot by day, faceted by condition
myTAI::plot_signature(example_phyex_set_sc, 
                     primary_identity = "day", 
                     secondary_identity = "batch",
                     facet_by_secondary = TRUE,
                     show_p_val = FALSE)

plot_signature single-cell with faceting

Other single-cell visualizations

The gene heatmap function also works with single-cell data and can show individual cells or be aggregated:

# Gene heatmap for single-cell data (aggregated by cell type)
myTAI::plot_gene_heatmap(example_phyex_set_sc, top_p = 0.1, cluster_rows=TRUE)

plot_gene_heatmap single-cell

# Gene heatmap showing individual cells (subsampled)
myTAI::plot_gene_heatmap(example_phyex_set_sc, show_reps = TRUE, max_cells_per_type = 10, top_p = 0.05, cluster_rows=TRUE)

plot_gene_heatmap single-cell with individual cells

# Change identity to "day" and plot heatmap grouped by developmental time
example_sc_by_day <- example_phyex_set_sc
example_sc_by_day@selected_idents <- "day"
myTAI::plot_gene_heatmap(example_sc_by_day, show_reps = TRUE, max_cells_per_type = 8, top_p = 0.05, cluster_rows=TRUE, show_gene_ids=TRUE, std=FALSE)

plot_gene_heatmap single-cell grouped by day

Single-cell plotting tips:

Use primary_identity to specify which metadata column to plot on the x-axis
Use secondary_identity with facet_by_secondary = TRUE for faceted plots
Use secondary_identity without faceting for colour-coded plots
Set custom colors with set_identity_colours()
Check available metadata columns with available_identities()

Plot away!