NEST - Teal

An overview of Teal project

Useful material

• Getting started with Teal
• Teal workshop PHUSE CSS 2025
• Teal gallery: sample apps based on teal framework
• Teal technical blueprint
• Video: a complete guide to getting started with teal. This video breaks down the components in a teal app, very useful

Running my first app

Mandatory arguments: data, module

Loads:

• shiny
• teal.data
• teal.code
• teal.slice

Main panel (Active Module Content)):

• shows two modules that loads IRIS data
• can have multiple

Side (right) (Filter Panel):

• active filter summary: number of observations
• active filter variables

Teal packages

• teal.data: defines data of teal app
• teal.slice: defines data filtering before passing into modules

And others

Creating a teal app

The bare bone structure is like this below:

library(teal)

app <- init(
    data = data, 
    modules = modules)

shinyApp(ui = app$ui, server = app$server)

Data

• Input: data.frame, cdisc data format data.frame, multi-asset objects
• teal_data(), if data is common across modules
• teal_data_module() if data is different

Modules

• teal.modules.general(): data.frame
• teal.modules.clinical(): cdisc data
• teal.modules.hermes(): multi-asset objects (check out TLG.catelog, biomarker.catelog)

Part 1: teal_data()

Creates an S4 object, <teal_data>

Slots: env, datanames, code, verified, join_keys (only useful for cdisc)

Methods: …

# create a single data object 
data <- teal_data(
  MTCARS = head(mtcars)
)

Pharmaverse

How does teal fit in the pharmaverse ecosystem? (Examples)

SDTM

ADaM

TLG

Calculation and analysis

• tern: analysis functions for tables and graphs in clinical trial reporting
• osprey:
• goshawk
• hermes

Interactive (teal)

Calculations above link to the teal modules to add interactivity. In addition to teal.moduels.general, we have

• tern -> teal.modules.clinical
• osprey -> teal.osprey
• goshawk -> teal.goshawk
• hermes -> teal.modules.hermes