loadL1K: loadL1K

In slinky: Putting the fun in LINCS L1000 data analysis

Description

Convert data from gctx file to SummarizedExperiment, pulling metadata from various sources. Specifying the subset of data you want can be done is various ways. The simplest example is with explicit subsetting (e.g. data <- loadL1K(sl[1:978, 1:10]), which is equivalent to data <- as(sl[1:978, 1:10], "SummarizedExperiment")). However, more sophisticated data loading can be achieved, for example by specifing a specific perturbagen (pert), or even an explicit where_clause which will be passed directly to the clue.io API. Thus, this function supports users with varying degrees of familiarity with the structure and content of the L1000 metadata. All arguments are optional.

Usage

loadL1K(x, ids = character(), pert = character(),
  cell_line = character(), type = NULL, where_clause = character(),
  gold = TRUE, inferred = TRUE, fields = character(),
  controls = FALSE, cl = NULL, verbose = FALSE)

## S4 method for signature 'Slinky'
loadL1K(x, ids = character(), pert = character(),
  cell_line = character(), type = NULL, where_clause = character(),
  gold = FALSE, inferred = TRUE, fields = character(),
  controls = FALSE, cl = NULL, verbose = FALSE)

Arguments

x	A slinky object
ids	distil_ids to include in the Expression Set.
pert	name (pert_iname) of perturbation for which data is desired. Supercedes value for pert_iname specified in where_clause (if any).
cell_line	name (cell_id) of cell line for which data is desired. Supercedes value for cell_id specified in where_clause (if any).
type	Optional type (pert_type) of perturbation for which data is desired, should be one of c("trt_cp", "trt_sh", "trt_oe") or NULL (default)
where_clause	Rather than specifying above terms, an explicit where_clause may be provided to identify the data to be loaded from the gctx file. This will be passed directly the the /pert endpoint of the clue.io API, and full documentation of the query options can be reviewed at the above link.
gold	Should we limit to instances classified as "gold" by the L1000 project (by virtue of their low inter-replicate variability)? Default is TRUE.
inferred	Should the inferred (non-landmark) genes be included in the analysis? Default is TRUE. Ignored if index is specified.
fields	Fields to include in the expression set's phenodata. Default is all available.
controls	Should same-plate controls be identified and included? Default is FALSE.
cl	Optional cluster object to speed up data retrieval from clue.io. Please use caution...a large cluster might produce requests to the API at an obnoxious rate.
verbose	Do you want to know how things are going? Default is FALSE

Value

Object of type SummarizedExperiment containing expression and meta data. #' @name loadL1K

Examples

# for build/demo only.  You MUST use your own key when using the slinky
# package.
user_key <- httr::content(httr::GET('https://api.clue.io/temp_api_key'),
                          as='parsed')$user_key
sl <- Slinky(user_key,
                 system.file('extdata', 'demo.gctx',
                      package='slinky'),
                 system.file('extdata', 'demo_inst_info.txt',
                     package = 'slinky'))
amox_gold <- clueInstances(sl, where_clause = list("pert_type" = "trt_cp",
                                                  "pert_iname" = "amoxicillin",
                                                  "cell_id" = "MCF7",
                                                  "is_gold" = TRUE), 
                          poscon = "omit")
ids.ctrl <- controls(sl, ids = amox_gold)$distil_id
amox_and_control <- loadL1K(sl, ids = c(amox_gold, 
                                       ids.ctrl))
str(SummarizedExperiment::assays(amox_and_control)[[1]]) 
                                      

slinky documentation built on Nov. 8, 2020, 10:58 p.m.