The Role of S4 in Bioconductor

Introduction to Bioconductor in R

James Chapman

Curriculum Manager, DataCamp

S3

Positive

CRAN, simple but powerful
Flexible and interactive
Uses a generic function
Functionality depends on the first argument
Example: plot() and methods(plot)

Negative

Bad at validating types and naming conventions (dot not dot?)
Inheritance works, but depends on the input

S4

Positive

Formal definition of classes
Bioconductor reusability
Has validation of types
Naming conventions

Example: myDescriptor <- new("GenomeDescription")

Negative

Complex structure compared to S3

Is it S4 or not?

Ask if an object is S4

isS4(myDescriptor)

TRUE

str of S4 objects start with Formal class

str(myDescriptor)

Formal class 'GenomeDescription' [package "GenomeInfoDb"] with 7 slots
...

S4 class definition

A class describes a representation

name
slots (methods/fields)
contains (inheritance definition)

MyEpicProject <- setClass(# Define class name with UpperCamelCase
                          "MyEpicProject",
                          # Define slots, helpful for validation
                          slots = c(ini = "Date", 
                                    end = "Date", 
                                    milestone = "character"),
                          # Define inheritance        
                          contains = "MyProject")

.S4methods(class = "GenomeDescription")

 [1] coerce          commonName      organism        provider       
 [5] providerVersion releaseDate     releaseName     seqinfo        
 [9] seqnames        show            toString        bsgenomeName   
see '?methods' for accessing help and source code

showMethods(classes = "GenomeDescription", where = search())

Object summary

show(myDescriptor)

| organism:  ()
| provider: 
| provider version: 
| release date: 
| release name: 
| ---
| seqlengths:

Let's practice!

Introduction to Bioconductor in R