--- title: "Beautiful plots made via myTAI" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{Beautiful plots made via myTAI} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- ## 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: → [📊](phylo-expression-object.html) ```{r message = FALSE, warning = FALSE} library(myTAI); library(S7); library(ggplot2); library(patchwork) data("example_phyex_set") ``` #### myTAI plots can be modified as a ggplot2 object. ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_signature function output with stat_flatline_test", dev.args = list(bg = 'transparent'), fig.align='center'} 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") ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_signature function output with stat_reductive_hourglass_test", dev.args = list(bg = 'transparent'), fig.align='center'} 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") ``` #### Transformation and robustness checks See more here: → [🛡️](tai-transform.html) ```{r message = FALSE, fig.height=5, fig.width=8, fig.alt="plot_signature_transformed function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_signature_transformed( example_phyex_set) ``` ```{r message = FALSE, fig.height=5, fig.width=8, fig.alt="plot_signature_gene_quantiles function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_signature_gene_quantiles( example_phyex_set) ``` #### Statistical tests and plotting results See more here: → [📈](tai-stats.html) ```{r message = FALSE, warning = FALSE, fig.height=3, fig.width=5, fig.alt="stat_flatline_test function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::stat_flatline_test( example_phyex_set, plot_result = TRUE) ``` ```{r message = FALSE, warning = FALSE} res_flt <- myTAI::stat_flatline_test(example_phyex_set, plot_result = FALSE) ``` ```{r message = FALSE, warning = FALSE, fig.height=5, fig.width=5, fig.alt="plot_cullen_frey function output for stat_flatline_test", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_cullen_frey(res_flt) ``` ```{r message = FALSE, warning = FALSE, fig.height=5, fig.width=5, fig.alt="plot_null_txi_sample function output for stat_flatline_test", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_null_txi_sample(res_flt) + ggplot2::guides(x = guide_axis(angle = 90)) ``` ```{r message = FALSE, warning = FALSE, fig.height=3, fig.width=5, fig.alt="stat_reductive_hourglass_test function output", dev.args = list(bg = 'transparent'), fig.align='center'} 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) ``` #### Average gene expression level by phylostratum ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_strata_expression function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_strata_expression(example_phyex_set) ``` `plot_strata_expression` with scaled y axis ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_strata_expression function output ggplot2", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_strata_expression(example_phyex_set) + ggplot2::scale_y_log10() + ggplot2::labs(x = "Expression aggregated by mean (log-scaled)") ``` `plot_strata_expression` with explicit transformation ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=8, fig.alt="plot_strata_expression function output", dev.args = list(bg = 'transparent'), fig.align='center'} 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 ``` 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 ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_contribution function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_contribution(example_phyex_set) ``` Curious about methods to obtain gene age information? See more here: → [📚](phylostratigraphy.html) For other analogous methods to assign evolutionary or expression information to each gene for TDI, TSI etc., see here: → [🧬](other-strata.html) ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=8, fig.alt="plot_distribution_expression function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_distribution_expression(example_phyex_set) ``` #### 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). ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=6, fig.width=9, fig.alt="plot_distribution_pTAI_qqplot function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_distribution_pTAI_qqplot(example_phyex_set) ``` #### Phylostratum distribution ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_distribution_strata function output", dev.args = list(bg = 'transparent'), fig.align='center'} 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)") ``` #### Expression heatmap ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=12, fig.width=12, fig.alt="plot_gene_heatmap function output default", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_gene_heatmap(example_phyex_set) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=12, fig.width=12, fig.alt="plot_gene_heatmap function output clustered", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_gene_heatmap(example_phyex_set, cluster_rows = TRUE, show_reps=TRUE, show_gene_ids=TRUE, top_p=0.005) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=12, fig.width=10, fig.alt="plot_gene_heatmap function output nonstd", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_gene_heatmap(example_phyex_set, cluster_rows = TRUE, show_reps=TRUE, top_p=0.005, std=FALSE, show_gene_ids=TRUE) ``` #### Dimension reduction ##### At the gene level ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_gene_space function output", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_gene_space(example_phyex_set) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_gene_space function output by strata", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_gene_space(example_phyex_set,colour_by = "strata") ``` ##### At the sample level ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=12, fig.alt="plot_sample_space function output by TXI", dev.args = list(bg = 'transparent'), fig.align='center'} myTAI::plot_sample_space(example_phyex_set) | myTAI::plot_sample_space(example_phyex_set, colour_by = "TXI") ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_sample_space function output by TXI", dev.args = list(bg = 'transparent'), fig.align='center', eval = requireNamespace("uwot", quietly = TRUE)} # we can even do a UMAP myTAI::plot_sample_space(example_phyex_set, method = "UMAP") ``` #### Inspecting mean-variance relationship ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=3, fig.width=8, fig.alt="plot_mean_var function output simple vs highlighted", dev.args = list(bg = 'transparent'), fig.align='center'} # 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") ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=3, fig.width=6, fig.alt="plot_gene_space function output log transform coloured by strata", dev.args = list(bg = 'transparent'), fig.align='center'} # 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)) ``` #### Individual gene expression profiles ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=12, fig.alt="plot_gene_profiles function output manual vs strata coloring", dev.args = list(bg = 'transparent'), fig.align='center'} # 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") ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_gene_profiles function output stage std_log transformation", dev.args = list(bg = 'transparent'), fig.align='center'} # stage colouring with standardized log transformation myTAI::plot_gene_profiles(example_phyex_set, max_genes = 10, transformation = "std_log", colour_by = "stage") ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=8, fig.width=12, fig.alt="plot_gene_profiles function output faceted by strata", dev.args = list(bg = 'transparent'), fig.align='center'} # 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) ``` ::: {.tip} 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: ```{r message = FALSE, warning = FALSE} # Load example single-cell data data(example_phyex_set_sc) ``` ```{r message = FALSE, warning = FALSE} example_phyex_set_sc ``` ```{r message = FALSE, warning = FALSE} # Check available identities cat("Available identities for plotting:\n") print(example_phyex_set_sc@available_idents) ``` ```{r message = FALSE, warning = FALSE} # 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 ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_signature single-cell basic", dev.args = list(bg = 'transparent'), fig.align='center'} # Basic signature plot showing TXI distribution across cell types myTAI::plot_signature(example_phyex_set_sc) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_signature single-cell without individual cells", dev.args = list(bg = 'transparent'), fig.align='center'} # Plot without showing individual cells (just means) myTAI::plot_signature(example_phyex_set_sc, show_reps = FALSE) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_signature single-cell by day", dev.args = list(bg = 'transparent'), fig.align='center'} # Plot TXI distribution by developmental day instead of cell type myTAI::plot_signature(example_phyex_set_sc, primary_identity = "day", show_p_val = FALSE) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=6, fig.alt="plot_signature single-cell by condition", dev.args = list(bg = 'transparent'), fig.align='center'} # Plot TXI distribution by experimental condition myTAI::plot_signature(example_phyex_set_sc, primary_identity = "condition", show_p_val=FALSE) ``` You can use a secondary identity for either coloring or faceting to create more informative plots: ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=4, fig.width=8, fig.alt="plot_signature single-cell with secondary coloring", dev.args = list(bg = 'transparent'), fig.align='center'} # Plot by day, colored by condition myTAI::plot_signature(example_phyex_set_sc, primary_identity = "day", secondary_identity = "condition", show_p_val=FALSE) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=6, fig.width=10, fig.alt="plot_signature single-cell with faceting", dev.args = list(bg = 'transparent'), fig.align='center'} # 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) ``` #### Other single-cell visualizations The gene heatmap function also works with single-cell data and can show individual cells or be aggregated: ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=8, fig.width=10, fig.alt="plot_gene_heatmap single-cell", dev.args = list(bg = 'transparent'), fig.align='center'} # 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) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=8, fig.width=12, fig.alt="plot_gene_heatmap single-cell with individual cells", dev.args = list(bg = 'transparent'), fig.align='center'} # 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) ``` ```{r message = FALSE, warning = FALSE, results = FALSE, fig.height=8, fig.width=12, fig.alt="plot_gene_heatmap single-cell grouped by day", dev.args = list(bg = 'transparent'), fig.align='center'} # 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) ``` ::: {.tip} **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!