Package 'plyxp'

Title: Data masks for SummarizedExperiment enabling dplyr-like manipulation
Description: The package provides `rlang` data masks for the SummarizedExperiment class. The enables the evaluation of unquoted expression in different contexts of the SummarizedExperiment object with optional access to other contexts. The goal for `plyxp` is for evaluation to feel like a data.frame object without ever needing to unwind to a rectangular data.frame.
Authors: Justin Landis [aut, cre] , Michael Love [aut]
Maintainer: Justin Landis <[email protected]>
License: MIT + file LICENSE
Version: 0.99.24
Built: 2024-10-29 03:29:04 UTC
Source: https://github.com/bioc/plyxp

Help Index


arrange rows or columns of PlySummarizedExperiment

Description

arrange() orders either the rows or columns of a PlySummarizedExperiment object. Note, to guarentee a valid PlySummarizedExperiment is returned, arranging in the assays evaluation context is disabled.

Unlike other dplyr verbs, arrange() largely ignores grouping. The PlySummarizedExperiment method also provides the same functionality via the .by_group argument.

Usage

## S3 method for class 'PlySummarizedExperiment'
arrange(.data, ..., .by_group = FALSE)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

...

<data-masking> Variables, or functions of variables. Use desc() to sort a variable in descending order.

.by_group

If TRUE, will sort first by grouping variable. Applies to grouped data frames only.

Value

an object inheriting PlySummarizedExperiment class

Examples

# arrange within rows/cols contexts separately
arrange(
  se_simple,
  rows(direction),
  cols(dplyr::desc(condition))
)

# access assay data to compute arrangement
arrange(
  se_simple,
  rows(rowSums(.assays_asis$counts)),
  cols(colSums(.assays_asis$counts))
)

# assay context is disabled
arrange(se_simple, counts) |> try()

# convert to `data.frame` first
as.data.frame(se_simple) |>
  arrange(counts)

create data.frame

Description

create data.frame

Usage

## S3 method for class 'PlySummarizedExperiment'
as.data.frame(x, ...)

Arguments

x

SummarizedExperiment object

...

unused arguments

Value

a data.frame object

Examples

as.data.frame(se_simple)

contextual plyxp pronouns

Description

plyxp utilizes its own version of rlang::.data pronouns. These may be used to gain access to other evaluation contexts for a managed set of data-masks.

Similar to rlang::.data, plyxp::.assays and other exported pronouns are exported to pass R CMD Checks. When using a plyxp within your package, import the associated pronoun from plyxp but only use the fully unqualified name, .assays, .assays_asis, etc.

Usage

.assays

.assays_asis

.rows

.rows_asis

.cols

.cols_asis

Format

An object of class NULL of length 0.

An object of class NULL of length 0.

An object of class NULL of length 0.

An object of class NULL of length 0.

An object of class NULL of length 0.

An object of class NULL of length 0.

Value

access to specific values behind the rlang pronoun

Examples

mutate(se_simple,
       # access via pronoun
       rows(sum = rowSums(.assays_asis$counts)),
       cols(sum = vapply(.assays$counts, sum, numeric(1))))

filter PlySummarizedExperiment

Description

The filter() function is used to subset an object, returning the observations that satisfy your conditions. An observation must return TRUE for all conditions within a context to be retained. Note, to guarantee a valid PlySummarizedExperiment is returned, filtering in the assays evaluation context is disabled.

Usage

## S3 method for class 'PlySummarizedExperiment'
filter(.data, ..., .preserve = FALSE)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

...

conditions to filter on. These must be wrapped in cols() and or rows()

.preserve

Relevant when the .data input is grouped. If .preserve = FALSE (the default), the grouping structure is recalculated based on the resulting data, i.e. the number of groups may change.

Value

an object inheriting PlySummarizedExperiment class

Examples

# example code
filter(
  se_simple,
  rows(length > 30),
  cols(condition == "drug")
)

filter(
  se_simple,
  rows(rowSums(.assays_asis$counts) > 40),
  cols(colSums(.assays_asis$counts) < 50)
)

# assay context is disabled
filter(
  se_simple,
  counts > 12
) |> try()

# convert to `data.frame` first
as.data.frame(se_simple) |>
  filter(counts > 12)

apply groups to PlySummarizedExperiment

Description

create grouping variables about the rowData and colData of a PlySummarizedExperiment object. Unlike the data.frame method the resulting output class is left unchanged. Thus dplyr generics for PlySummarizedExperiment must check grouping information manually.

Usage

## S3 method for class 'PlySummarizedExperiment'
group_by(.data, ..., .add = FALSE)

## S3 method for class 'PlySummarizedExperiment'
ungroup(x, ...)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

S4 Compatibility

At the moment, grouping on S4 Vectors is not yet supported. This is due to plyxp using ⁠[vec_group_loc][vctrs::vec_group_loc]⁠ to form grouping information. plyxp will eventually develop a method to handle S4 Vectors.

...

contextual expressions specifying which columns to ungroup. Omitting ... ungroups the entire object.

.add

When FALSE, the default, group_by() will override existing groups.

x

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

Value

PlySummarizedExperiment object

Functions

  • ungroup(PlySummarizedExperiment): Ungroup a PlySummarizedExperiment object

Examples

group_by(se_simple, rows(direction), cols(condition))

get grouping data

Description

retrieve grouping information from a SummarizedExperiment object. This is stored within the metadata() of the object.

Usage

## S3 method for class 'PlySummarizedExperiment'
group_data(.data)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

Value

list of groupings for an SummarizedExperiment

Examples

group_by(se_simple, rows(direction), cols(condition)) |> group_data()

get PlySummarizedExperiment grouping Variables

Description

like in dplyr::group_vars() will get character strings for groupings with the exection of the return value being a list for each grouped context

Usage

## S3 method for class 'PlySummarizedExperiment'
group_vars(x)

Arguments

x

PlySummarizedExperiment

Value

NULL or list containing names of grouping columns

Examples

out <- group_by(se_simple, rows(direction))
group_vars(out)

Mutate a PlySummarizedExperiment object

Description

Mutate a PlySummarizedExperiment object under an data mask. Unlike a few other dplyr implementations, all contextual evaluations of mutate() for SummarizedExperiment are valid.

Usage

## S3 method for class 'PlySummarizedExperiment'
mutate(.data, ...)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

...

expressions to evaluate

Value

an object inheriting PlySummarizedExperiment class

Examples

mutate(se_simple,
  counts_1 = counts + 1,
  logp_counts = log(counts_1),
  # access assays context with ".assays" pronoun,
  # note that assays are sliced into a list to
  # fit dimensions of cols context
  cols(sum = purrr::map_dbl(.assays$counts, sum)),
  # access assays context "asis" with the same pronoun
  # but with a "_asis" suffix.
  rows(sum = rowSums(.assays_asis$counts))
)

SummarizedExperiment Shell Object

Description

A container object for the SummarizedExperiment class.

This S4 class is implemented to bring unique dplyr syntax to the SummarizedExperiment object without clashing with the tidySummarizedExperiment package. As such, this is a simple wrapper that contains one slot, which holds a SummarizedExperiment object.

Usage

new_plyxp(se)

PlySummarizedExperiment(se)

Arguments

se

SummarizedExperiment object

Value

PlySummarizedExperiment object

Slots

se

contains the underlying SummarizedExperiment class.

Examples

se <- SummarizedExperiment(
  assays = list(counts = matrix(1:6, nrow = 3)),
  colData = S4Vectors::DataFrame(condition = c("A", "B"))
)
new_plyxp(se = se)
# or
PlySummarizedExperiment(se = se)

Modify SummarizedExperiment Object

Description

Modify the underlying SummarizedExperiment object with a function.

Usage

plyxp(.data, .f, ...)

Arguments

.data

a PlySummarizedExperiment object

.f

a function that returns a SummarizedExperiment object

...

additional arguments passed to .f

Value

a PlySummarizedExperiment object

Examples

plyxp(se_simple, function(x) x)

plyxp contexts

Description

Contextual user-facing helper function for dplyr verbs with SummarizedExperiment objects. These functions are intended to be used as the top level call to any dplyr verbs ... argument, similar to that of across()/if_any()/if_all().

Specifies that the following expressions should be evaluated within the colData context.

Specifies that the following expressions should be evaluated within the rowData context.

Specify a single expression to evaluate in another context

Specify a single expression to evaluate in another context

Specify a single expression to evaluate in another context

Usage

cols(...)

rows(...)

col_ctx(x, asis = FALSE)

row_ctx(x, asis = FALSE)

assay_ctx(x, asis = FALSE)

Arguments

x, ...

expressions to evaluate within its associated context

asis

asis = FALSE (the default) will indicate using active bindings that attempt to coerce the underlying data into a format that is appropriate for the current context. Indicating TRUE will instead bind the underlying data as is.

Value

function called for its side-effects

Examples

# cols
mutate(se_simple,
       cols(is_drug = condition=="drug"),
       #bind a different context
       effect = col_ctx(counts + (is_drug * rbinom(n(), 20, .3))))

extract data from object

Description

similar to dplyr::pull.data.frame except allows to extract objects from different contexts.

Usage

## S3 method for class 'PlySummarizedExperiment'
pull(.data, var = -1, name = NULL, ...)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

var

A variable as specified by dplyr::pull

name

ignored argument. Due to the range of data types a PlySummarizedExperiment this argument is not supported

...

unused argument

Value

an element from either the assays, rowData, or colData of a SummarizedExperiment object

Examples

# last element of default context (assays)
pull(se_simple, var = -1)
# first element of rows context
pull(se_simple, var = rows(1))
# element from col context by literal variable name
pull(se_simple, var = cols(condition))

# use `pull()` to return contextual info
mutate(se_simple, rows(counts = .assays$counts)) |>
  # get last stored element
  pull(rows(-1))

PlySummarizedExperiment Methods

Description

Methods from SummarizedExperiment package re-implemented for PlySummarizedExperiment.

Usage

se(x)

## S4 method for signature 'PlySummarizedExperiment'
se(x)

se(x) <- value

## S4 replacement method for signature 'PlySummarizedExperiment'
se(x) <- value

## S4 method for signature 'PlySummarizedExperiment'
assays(x, withDimnames = TRUE, ...)

## S4 replacement method for signature 'PlySummarizedExperiment,list'
assays(x, withDimnames = TRUE, ...) <- value

## S4 replacement method for signature 'PlySummarizedExperiment,SimpleList'
assays(x, withDimnames = TRUE, ...) <- value

## S4 method for signature 'PlySummarizedExperiment,missing'
assay(x, i, withDimnames = TRUE, ...)

## S4 method for signature 'PlySummarizedExperiment,numeric'
assay(x, i, withDimnames = TRUE, ...)

## S4 method for signature 'PlySummarizedExperiment,character'
assay(x, i, withDimnames = TRUE, ...)

## S4 replacement method for signature 'PlySummarizedExperiment,missing'
assay(x, i, withDimnames = TRUE, ...) <- value

## S4 replacement method for signature 'PlySummarizedExperiment,numeric'
assay(x, i, withDimnames = TRUE, ...) <- value

## S4 replacement method for signature 'PlySummarizedExperiment,character'
assay(x, i, withDimnames = TRUE, ...) <- value

## S4 method for signature 'PlySummarizedExperiment'
rowData(x, use.names = TRUE, ...)

## S4 replacement method for signature 'PlySummarizedExperiment'
rowData(x, ...) <- value

## S4 method for signature 'PlySummarizedExperiment'
colData(x, ...)

## S4 replacement method for signature 'PlySummarizedExperiment,DataFrame'
colData(x, ...) <- value

## S4 replacement method for signature 'PlySummarizedExperiment,NULL'
colData(x, ...) <- value

Arguments

x

PlySummarizedExperiment object

value

replacement value

withDimnames

logical

...

additional arguments

i

character or numeric index

use.names

logical

Value

Replacement functions return a PlySummarizedExperiment object. Other functions will return the same object as the method from SummarizedExperiment.

Functions

  • se(PlySummarizedExperiment): get the se slot of the PlySummarizedExperiment object

  • se(x) <- value: set the se slot of the PlySummarizedExperiment object

  • se(PlySummarizedExperiment) <- value: set the se slot of the PlySummarizedExperiment object

  • assays(PlySummarizedExperiment): get the assays o the PlySummarizedExperiment object

  • assays(x = PlySummarizedExperiment) <- value: set the assays of the PlySummarizedExperiment object

  • assays(x = PlySummarizedExperiment) <- value: set the assays of the PlySummarizedExperiment object

  • assay(x = PlySummarizedExperiment, i = missing): get the first assay of the PlySummarizedExperiment object

  • assay(x = PlySummarizedExperiment, i = numeric): get assay from a PlySummarizedExperiment object

  • assay(x = PlySummarizedExperiment, i = character): get assay from a PlySummarizedExperiment object

  • assay(x = PlySummarizedExperiment, i = missing) <- value: set assay in a PlySummarizedExperiment object

  • assay(x = PlySummarizedExperiment, i = numeric) <- value: set assay in a PlySummarizedExperiment object

  • assay(x = PlySummarizedExperiment, i = character) <- value: set assay in a PlySummarizedExperiment object

  • rowData(PlySummarizedExperiment): get rowData in a PlySummarizedExperiment object

  • rowData(PlySummarizedExperiment) <- value: set rowData in a PlySummarizedExperiment object

  • colData(PlySummarizedExperiment): get colData in a PlySummarizedExperiment object

  • colData(x = PlySummarizedExperiment) <- value: set colData in a PlySummarizedExperiment object

Examples

assays(se_simple)
rowData(se_simple)
colData(se_simple)

Plyxp Simple Example Summarized Experiment

Description

A small data SummarizedExperiment Object of 20 observations, 5 rows and 4 columns.

Usage

se_simple

Format

se_simple

assays
counts

sampled data points between 1:20

logcounts

log transform of counts

rowData/.features
gene

fake gene name

length

fake gene length

direction

fake strand

colData/.samples
sample

fake sample name

condition

control or drug treatment

Value

a SummarizedExperiment object

Examples

SummarizedExperiment::assays(se_simple)
SummarizedExperiment::rowData(se_simple)
SummarizedExperiment::colData(se_simple)

select assays, rowData, and colData names

Description

Select one or more values from each context. By default omitting an expression for a context is the same as selecting NOTHING from that context.

The <tidy-select> implementation within plyxp is almost similar to dplyr except when used within the across() function. When used from accross(), the data provided to eval_select is a zero length slice of the data. This was an intentional choice to prevent the evaluation of potentionally expensive chopping operations for S4Vectors. This means that predicate function from where() will NOT be able to query the original data.

Usage

## S3 method for class 'PlySummarizedExperiment'
select(.data, ...)

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

...

<tidy-select> one or more selection expressions. Supports wrapping expressions within the <plyxp-contexts>.

Value

an object inheriting PlySummarizedExperiment class

Examples

# only keep assays, other contexts are dropped
select(se_simple, everything())

# only keep rowData, other contexts are dropped
select(se_simple, rows(everything()))

select(se_simple, rows(where(is.numeric)))

# Note on `where()` clause, all data is available within select
select(se_simple, rows(where(~any(grepl("-", .x)))))

# within an `across()`, only a zero-length slice avialble, so the
# `where()` predicate cannot access the data
mutate(se_simple,
       rows(
        across(where(~any(grepl("-", .x))),
               ~sprintf("%s foo", .x))))
# here is an acceptable usage of the `where()` predicate
mutate(se_simple,
       rows(
        across(where(is.character),
               ~sprintf("%s foo", .x))))

Summarize PlySummarizedExperiment

Description

Summarize PlySummarizedExperiment

Usage

## S3 method for class 'PlySummarizedExperiment'
summarize(.data, ..., .retain = c("auto", "ungrouped", "none"))

## S3 method for class 'PlySummarizedExperiment'
summarise(.data, ..., .retain = c("auto", "ungrouped", "none"))

Arguments

.data

An object Inheriting from PlySummarizedExperiment, the wrapper class for SummarizedExperiment objects

...

expressions to summarize the object

.retain

This argument controls how rowData() or colData() is retained after summarizing. When "auto" (the default), .retain behavior depends on the groupings of .data. When exactly one dimension is grouped, "auto" behaves like "ungrouped-dim", and "none" otherwise. When "ungrouped-dim", the ungrouped dimension's data are retained in the resulting SummarizedExperiment object and scalar outputs are recycled to the length of the ungrouped dimension. When "none", all outputs are expected to be scalar and only computed values are retained in rowData() and colData()

Value

an object inheriting PlySummarizedExperiment class

Examples

# outputs in assay context may be either
# length 1, or the length of the ungrouped
# dimension while .retain = "auto"/"ungrouped-dim"
se_simple |>
  group_by(rows(direction)) |>
  summarise(
    col_sums = colSums(counts),
    sample = sample(1:20, 1L)
  )

# .retain = "none" will drop ungrouped dimensions and
# outputs of assay context should be length 1.
se_simple |>
  group_by(rows(direction)) |>
  summarise(
    col_sums = list(colSums(counts)),
    .retain = "none"
  )

# using an `across()` function will help
# nest ungrouped dimensions
se_simple |>
  group_by(rows(direction)) |>
  summarise(
    col_sums = list(colSums(counts)),
    cols(across(everything(), list)),
    .retain = "none"
  )

Get observations of a vector

Description

This extends vctrs::vec_slice to S4Vectors::Vector class by masking vec_slice with S7::new_generic. Atomic vectors and other base S3 classes (list, data.frame, factor, Dat, POSIXct) will dispatch to the vctrs::vec_slice method as normal. Dispatch support on the S4Vectors::Vector and S4Vectors::DataFrame classes provides a unified framework for working with base R vectors and S4Vectors.

S4Vectors::Vector Implementation

This method will naively call the [ method for any S4 class that inherits from the S4Vectors::Vector class. This may not be a very efficient way to slice up an S4 class, but will work.

With this implementation, the x@mcol data is expected to be retained after a call to plyxp::vec_slice(x, i).

S4Vectors::DataFrame Implementation

The DataFrame implementation works similar to how vctrs::vec_slice works on a data.frame object. What is being sliced is the rows of x@listData. To maintain the size stability of the DataFrame object, we change ⁠@nrows⁠ to the appropriate value, and perform a recursive call if ⁠@elementMetadata⁠ is not NULL.

Performance

Depending on the size and complexity of your S4 Vector object, you may find the standard subset operation is extremely slow. For example, consider a SummarizedExperiment whose rowData contains a CompressedGRangesList object assigned to the name "exons" and whose length is 250,000 and underlying ⁠@unlistData⁠ is length 1,600,000. Performing a by .features grouping operation and attempting to evaluate the exons within the row context would force the CompressedGRangesList object to be chopped element-wise.

Unfortunately, there is a massive performance hit in attempting to construct 250,000 GRanges. Unless you do not mind waiting over an hour for each dplyr verb in which exons gets evaluated, doing so is not recommended.

The plyxp package is planning to export a new generic named plyxp_s4_proxy_vec(). This attempts to reconstruct certain standard S4Vectors::Vectors as standard vectors or tibbles. The equivalent exons object would require much more memory use, but at the advantage of only taking several seconds to construct. When you are done, you can attempt to restore the original S4 Vector with plyxp_restore_s4_proxy().

In development, plyxp_s4_proxy_vec() is faster to work with because there are less checks on the object validity and all ⁠@elementMetadata⁠ and ⁠@metadata⁠ are dropped from the objects.

Usage

vec_slice(x, i, ...)

Arguments

x

A vector

i

An integer, character or logical vector specifying the locations or names of the observations to get/set. Specify TRUE to index all elements (as in x[]), or NULL, FALSE or integer() to index none (as in x[NULL]).

...

These dots are for future extensions and must be empty.

Value

a new S3 or S4 vector subsetted by i

Examples

vec_slice(1:10, i = 5)
vec_slice(S4Vectors::Rle(rep(1:3, each = 3)), i = 5)

Recycle a vector

Description

A re-export of vctrs::vec_recycle as an S7 generic function to allow S4Vectors.

Usage

vec_recycle(x, size, ..., x_arg = "", call = caller_env())

Arguments

x

A vector to recycle.

size

Desired output size.

...

Depending on the function used:

  • For vec_recycle_common(), vectors to recycle.

  • For vec_recycle(), these dots should be empty.

x_arg

Argument name for x. These are used in error messages to inform the user about which argument has an incompatible size.

call

The execution environment of a currently running function, e.g. caller_env(). The function will be mentioned in error messages as the source of the error. See the call argument of abort() for more information.

Value

a S3 or S4 vector

Examples

vec_recycle(1L, size = 5L)
vec_recycle(S4Vectors::Rle(1L), size = 5L)

replicate a vector

Description

A re-export of vctrs::vec_rep and vctrs::vec_rep_each as an S7 generic function to allow S4Vectors.

Usage

vec_rep(
  x,
  times,
  ...,
  error_call = caller_env(),
  x_arg = "x",
  times_arg = "times"
)

vec_rep_each(
  x,
  times,
  ...,
  error_call = caller_env(),
  x_arg = "x",
  times_arg = "times"
)

Arguments

x

A vector.

times

For vec_rep(), a single integer for the number of times to repeat the entire vector.

For vec_rep_each(), an integer vector of the number of times to repeat each element of x. times will be recycled to the size of x.

...

These dots are for future extensions and must be empty.

error_call

The execution environment of a currently running function, e.g. caller_env(). The function will be mentioned in error messages as the source of the error. See the call argument of abort() for more information.

x_arg, times_arg

Argument names for errors.

Value

a new S3 or S4 vector replicated by specified times

Examples

vec_rep(1:2, times = 5)
vec_rep(S4Vectors::Rle(1:2), times = 5)

vec_rep_each(1:2, times = 5)
vec_rep_each(S4Vectors::Rle(1:2), times = 5)

Printing within tibble with S4 objects

Description

plyxp uses pillar for its printing. If you want to change how your S4 object is printed within plyxp's print method, consider writing a method for this function.

To print S4 objects in a tibble, plyxp hacks a custom integer vector built from vctrs where the S4 object lives in an attribute named "phantomData". You can create your own S4 phantom vector with vec_phantom(). This function is not used outside of printing for plyxp

The default method for formatting a vec_phantom() is to call showAsCell().

Usage

vec_phantom(x)

plyxp_pillar_format(x, ...)

show_tidy(x, ...)

use_show_tidy()

use_show_default()

Arguments

x

The S4 object

...

other arguments passed from pillar_shaft

Value

plyxp_pillar_format -> formatted version of your S4 vector vec_phantom -> integer vector with arbitrary object in phatomData attribute.

tidy printing

By default, plyxp will not affect the show method for SummarizedExperiment objects. In order to use a tibble abstraction, use use_show_tidy() to enable or use_show_default() to disable this feature. These functions are called for their side effects, modifying the global option "show_SummarizedExperiment_as_tibble_abstraction".

To show an object as the tibble abstraction regardless of the set option, use the S3 generic show_tidy(...).

Examples

if(require("IRanges")) {
  ilist <- IRanges::IntegerList(list(c(1L,2L,3L),c(5L,6L)))
  phantom <- vec_phantom(ilist)
  pillar::pillar_shaft(phantom)
  
  plyxp_pillar_format.CompressedIntegerList <- function(x) {
   sprintf("Int: [%i]", lengths(x))
  }
  pillar::pillar_shaft(phantom)
  rm(plyxp_pillar_format.CompressedIntegerList)
}

# default printing
se_simple
# use `plyxp` tibble abstraction
use_show_tidy()
se_simple
# restore default print
use_show_default()
se_simple
# explicitly using tibble abstraction
show_tidy(se_simple)

S7 classes for vctrs and S4 Vectors

Description

A set of S7 classes and Class unions that help establish S7 method dispatch. These classes were made to re-export several vctrs functions such that internals for plyxp were consistent with room for optimization.

Usage

class_vctrs

class_s4_vctrs

class_DF

Format

An object of class S7_union of length 1.

An object of class classRepresentation of length 1.

An object of class classRepresentation of length 1.

Value

S7 class union or base class

See Also

vec_rep(),vec_recycle(),vec_slice()