An Overview of the S4Vectors package (original) (raw)
In the context of the S4Vectors package, a vector-like object is an ordered finite collection of elements. All vector-like objects have three main properties: (1) a notion of length or number of elements, (2) the ability to extract elements to create new vector-like objects, and (3) the ability to be concatenated with one or more vector-like objects to form larger vector-like objects. The main functions for these three operations are length, [, andc. Supporting these operations provide a great deal of power and many vector-like object manipulations can be constructed using them.
Some vector-like objects can also have a list-like semantic, which means that individual elements can be extracted with [[.
In S4Vectors and many other Bioconductor packages, vector-like and list-like objects derive from the Vector and List virtual classes, respectively. Note that List is a subclass of Vector. The following subsections describe each in turn.
Vector-like objects
As a first example of vector-like objects, we’ll look at Rle objects. In R, atomic sequences are typically stored in atomic vectors. But there are times when these object become too large to manage in memory. When there are lots of consecutive repeats in the sequence, the data can be compressed and managed in memory through a run-length encoding where a data value is paired with a run length. For example, the sequence {1, 1, 1, 2, 3, 3} can be represented as values = {1, 2, 3}, run lengths = {3, 1, 2}.
The Rle class defined in the S4Vectors package is used to represent a run-length encoded (compressed) sequence of logical, integer,numeric, complex, character, raw, or factor values. Note that the_Rle_ class extends the Vector virtual class:
showClass("Rle")## Class "Rle" [package "S4Vectors"]
##
## Slots:
##
## Name: values lengths elementMetadata metadata
## Class: vector_OR_factor integer_OR_LLint DataFrame_OR_NULL list
##
## Extends:
## Class "Vector", directly
## Class "Annotated", by class "Vector", distance 2
## Class "vector_OR_Vector", by class "Vector", distance 2One way to construct Rle objects is through the Rle constructor function:
set.seed(0)
lambda <- c(rep(0.001, 4500), seq(0.001, 10, length=500),
seq(10, 0.001, length=500))
xVector <- rpois(1e7, lambda)
yVector <- rpois(1e7, lambda[c(251:length(lambda), 1:250)])
xRle <- Rle(xVector)
yRle <- Rle(yVector)Rle objects are vector-like objects:
length(xRle)## [1] 10000000xRle[1]## integer-Rle of length 1 with 1 run
## Lengths: 1
## Values : 0zRle <- c(xRle, yRle)Subsetting a vector-like object
As with ordinary R atomic vectors, it is often necessary to subset one sequence from another. When this subsetting does not duplicate or reorder the elements being extracted, the result is called a subsequence. In general, the[ function can be used to construct a new sequence or extract a subsequence, but its interface is often inconvenient and not amenable to optimization. To compensate for this, the S4Vectors package supports seven additional functions for sequence extraction:
1.window - Extracts a subsequence over a specified region.
2.subset - Extracts the subsequence specified by a logical vector.
3.head - Extracts a consecutive subsequence containing the first n elements.
4.tail - Extracts a consecutive subsequence containing the last n elements.
5.rev - Creates a new sequence with the elements in the reverse order.
6.rep - Creates a new sequence by repeating sequence elements.
The following code illustrates how these functions are used on an Rle vector:
xSnippet <- window(xRle, 4751, 4760)
xSnippet## integer-Rle of length 10 with 9 runs
## Lengths: 1 1 1 1 1 1 1 1 2
## Values : 4 6 5 4 6 2 6 7 5head(xSnippet)## integer-Rle of length 6 with 6 runs
## Lengths: 1 1 1 1 1 1
## Values : 4 6 5 4 6 2tail(xSnippet)## integer-Rle of length 6 with 5 runs
## Lengths: 1 1 1 1 2
## Values : 6 2 6 7 5rev(xSnippet)## integer-Rle of length 10 with 9 runs
## Lengths: 2 1 1 1 1 1 1 1 1
## Values : 5 7 6 2 6 4 5 6 4rep(xSnippet, 2)## integer-Rle of length 20 with 18 runs
## Lengths: 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 2
## Values : 4 6 5 4 6 2 6 7 5 4 6 5 4 6 2 6 7 5subset(xSnippet, xSnippet >= 5L)## integer-Rle of length 7 with 5 runs
## Lengths: 1 1 2 1 2
## Values : 6 5 6 7 5Concatenating vector-like objects
The S4Vectors package uses two generic functions, c andappend, for concatenating two Vector derivatives. The methods for Vector_objects follow the definition that these two functions are given the_base package.
c(xSnippet, rev(xSnippet))## integer-Rle of length 20 with 17 runs
## Lengths: 1 1 1 1 1 1 1 1 4 1 1 1 1 1 1 1 1
## Values : 4 6 5 4 6 2 6 7 5 7 6 2 6 4 5 6 4append(xSnippet, xSnippet, after=3)## integer-Rle of length 20 with 18 runs
## Lengths: 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 2
## Values : 4 6 5 4 6 5 4 6 2 6 7 5 4 6 2 6 7 5Looping over subsequences of vector-like objects
In R, for looping can be an expensive operation. To compensate for this, the_S4Vectors_ package provides aggregate and shiftApply methods (shiftApply is a new generic function defined in S4Vectors to perform calculations over subsequences of vector-like objects.
The aggregate function combines sequence extraction functionality of thewindow function with looping capabilities of the sapply function. For example, here is some code to compute medians across a moving window of width 3 using the function aggregate:
xSnippet## integer-Rle of length 10 with 9 runs
## Lengths: 1 1 1 1 1 1 1 1 2
## Values : 4 6 5 4 6 2 6 7 5aggregate(xSnippet, start=1:8, width=3, FUN=median)## [1] 5 5 5 4 6 6 6 5The shiftApply function is a looping operation involving two vector-like objects whose elements are lined up via a positional shift operation. For example, the elements of xRle and yRle were simulated from Poisson distributions with the mean of element i from yRle being equivalent to the mean of element i + 250 from xRle. If we did not know the size of the shift, we could estimate it by finding the shift that maximizes the correlation betweenxRle and yRle.
cor(xRle, yRle)## [1] 0.5739224shifts <- seq(235, 265, by=3)
corrs <- shiftApply(shifts, yRle, xRle, FUN=cor)plot(shifts, corrs)Figure 1: Correlation between xRle and yRle for various shifts
The result is shown in Fig.1
More on Rle objects
When there are lots of consecutive repeats, the memory savings through an RLE can be quite dramatic. For example, the xRle object occupies less than one third of the space of the original xVector object, while storing the same information:
as.vector(object.size(xRle) / object.size(xVector))## [1] 0.3020726identical(as.vector(xRle), xVector)## [1] TRUEThe functions runValue and runLength extract the run values and run lengths from an Rle object respectively:
head(runValue(xRle))## [1] 0 1 0 1 0 1head(runLength(xRle))## [1] 780 1 208 1 1599 1The Rle class supports many of the basic methods associated with R atomic vectors including the Ops, Math, Math2, Summary, and Complex group generics. Here is a example of manipulating Rle objects using methods from the Ops group:
xRle > 0## logical-Rle of length 10000000 with 197127 runs
## Lengths: 780 1 208 1 1599 ... 5 1 91 1 927
## Values : FALSE TRUE FALSE TRUE FALSE ... FALSE TRUE FALSE TRUE FALSExRle + yRle## integer-Rle of length 10000000 with 1957707 runs
## Lengths: 780 1 208 1 13 1 413 1 ... 1 5 1 91 1 507 1 419
## Values : 0 1 0 1 0 1 0 1 ... 2 0 1 0 1 0 1 0xRle > 0 | yRle > 0## logical-Rle of length 10000000 with 210711 runs
## Lengths: 780 1 208 1 13 ... 91 1 507 1 419
## Values : FALSE TRUE FALSE TRUE FALSE ... FALSE TRUE FALSE TRUE FALSEHere are some from the Summary group:
range(xRle)## [1] 0 26sum(xRle > 0 | yRle > 0)## [1] 2105185And here is one from the Math group:
log1p(xRle)## numeric-Rle of length 10000000 with 1510219 runs
## Lengths: 780 1 208 1 ... 91 1 927
## Values : 0.000000 0.693147 0.000000 0.693147 ... 0.000000 0.693147 0.000000As with atomic vectors, the cor and shiftApply functions operate on _Rle_objects:
cor(xRle, yRle)## [1] 0.5739224shiftApply(249:251, yRle, xRle,
FUN=function(x, y) {var(x, y) / (sd(x) * sd(y))})## [1] 0.8519138 0.8517324 0.8517725For more information on the methods supported by the Rle class, consult theRle man page.
List-like objects
Just as with ordinary R List objects, _List_-derived objects support [[for element extraction, c for concatenating, and lapply/sapply for looping. lapply and sapply are familiar to many R users since they are the standard functions for looping over the elements of an R list object.
In addition, the S4Vectors package introduces the endoapplyfunction to perform an endomorphism equivalent to lapply, i.e. it returns a_List_ derivative of the same class as the input rather than a list object.
An example of List derivative is the DataFrame class:
showClass("DataFrame")## Virtual Class "DataFrame" [package "S4Vectors"]
##
## Slots:
##
## Name: elementType elementMetadata metadata
## Class: character DataFrame_OR_NULL list
##
## Extends:
## Class "RectangularData", directly
## Class "List", directly
## Class "DataFrame_OR_NULL", directly
## Class "Vector", by class "List", distance 2
## Class "list_OR_List", by class "List", distance 2
## Class "Annotated", by class "List", distance 3
## Class "vector_OR_Vector", by class "List", distance 3
##
## Known Subclasses: "DFrame"One way to construct DataFrame objects is through the DataFrame constructor function:
df <- DataFrame(x=xRle, y=yRle)
sapply(df, class)## x y
## "Rle" "Rle"sapply(df, summary)## x y
## Min. 0.0000000 0.0000000
## 1st Qu. 0.0000000 0.0000000
## Median 0.0000000 0.0000000
## Mean 0.9090338 0.9096009
## 3rd Qu. 0.0000000 0.0000000
## Max. 26.0000000 27.0000000sapply(as.data.frame(df), summary)## x y
## Min. 0.0000000 0.0000000
## 1st Qu. 0.0000000 0.0000000
## Median 0.0000000 0.0000000
## Mean 0.9090338 0.9096009
## 3rd Qu. 0.0000000 0.0000000
## Max. 26.0000000 27.0000000endoapply(df, `+`, 0.5)## DataFrame with 10000000 rows and 2 columns
## x y
## <Rle> <Rle>
## 1 0.5 0.5
## 2 0.5 0.5
## 3 0.5 0.5
## 4 0.5 0.5
## 5 0.5 0.5
## ... ... ...
## 9999996 0.5 0.5
## 9999997 0.5 0.5
## 9999998 0.5 0.5
## 9999999 0.5 0.5
## 10000000 0.5 0.5For more information on DataFrame objects, consult the DataFrame man page.
See the "An Overview of the IRanges package’’ vignette in the_IRanges_ package for many more examples of List derivatives.