bioshmods

bioshmods provides reusable Shiny modules for different tasks in exploration of different bioinformatics data, such as differential gene expression results and other high-throughput datasets. The modules include interactive tables, plots, and data export functionality to facilitate creating of custom bioinformatic Shiny apps.

Included modules

• geneBrowserTableServer() / geneBrowserPlotServer() for table-first gene browsing and expression profiling.
• volcanoServer() and discoServer() for interactive contrast visualization.
• pcaServer() for PCA exploration.
• tmodBrowserTableServer(), tmodBrowserPlotServer(), and tmodPanelPlotServer() for tmod gene-set results.
• geneGroupSelectorServer() to build gene lists by name, expression, or differential-expression criteria.
• fileExportServer() to export data frames and R objects (xlsx, rds, zip).

Installation

Install the development version from GitHub:

# install.packages("pak")
pak::pkg_install("bihealth/bioshmods")

Alternative:

# install.packages("remotes")
remotes::install_github("bihealth/bioshmods")

Example data

The package ships with C19, a list containing differential-expression contrasts, annotation, expression, and covariate data.

library(bioshmods)
data(C19)
names(C19)
#> [1] "contrasts"  "annotation" "expression" "covariates"
names(C19$contrasts)
#> [1] "COVID19_ID0" "ICU_ID1"
dim(C19$expression)
#> [1] 558  32
head(colnames(C19$annotation))
#> [1] "PrimaryID" "ENSEMBL"   "SYMBOL"    "ENTREZID"  "REFSEQ"    "GENENAME"

Quick start

Launch the integrated example gene browser:

library(bioshmods)
data(C19)

annot_linkout <- list(
  SYMBOL = "https://www.genecards.org/cgi-bin/carddisp.pl?gene=%s",
  ENTREZID = "https://www.ncbi.nlm.nih.gov/gene/?term=%s"
)

gene_browser(C19, annot_linkout = annot_linkout)

Data expectations

Many of the modules use similar data structures, for example:

• contrasts: named list of data frames; each contrast should include PrimaryID, log2FoldChange, and a p-value column (pvalue or padj, depending on module).
• annotation: data frame containing at least PrimaryID.
• expression: expression matrix/data frame with gene IDs in row names.
• covariates: sample metadata aligned to expression columns.

Communication between modules

Modules communicate through shared reactive objects. The general pattern is:

• create one reactiveValues() object in the app server
• pass it to a “producer” module (writes selected IDs)
• pass the same object to a “consumer” module (reacts to changes)

For example, in the gene browser, geneBrowserTableServer() writes gene_id$id and gene_id$ds, and geneBrowserPlotServer() observes those values and updates the plot accordingly.

library(shiny)
library(bioshmods)
data(C19)

# creeate the UI
ui <- fluidPage(
  fluidRow(
    column(6, geneBrowserTableUI("tbl", names(C19$contrasts))),
    column(6, geneBrowserPlotUI("gplot", contrasts = TRUE))
  )
)

# create the server
server <- function(input, output, session) {

  # this is the reactive value that will be 
  # used for communication between the table and plot modules
  gene_id <- reactiveValues()

  geneBrowserTableServer(
    "tbl",
    cntr = C19$contrasts,
    annot = C19$annotation,
    gene_id = gene_id
  )

  geneBrowserPlotServer(
    "gplot",
    gene_id = gene_id,
    covar = C19$covariates,
    exprs = C19$expression,
    annot = C19$annotation,
    cntr = C19$contrasts
  )
}

shinyApp(ui, server)

The same approach works with volcanoServer() and discoServer(), which can also publish selected genes into the same shared gene_id object.

Minimal module app

ui <- shiny::fluidPage(
  bioshmods::volcanoUI("volc", datasets = names(C19$contrasts))
)

server <- function(input, output, session) {
  bioshmods::volcanoServer(
    "volc",
    cntr = C19$contrasts,
    annot = C19$annotation
  )
}

shiny::shinyApp(ui, server)