Build a SummarizedExperiment from a long table

Turns a long table, one row per measurement, into an SE with a single BumpyMatrix assay. Meant for assays that do not come out of the gDR pipeline, e.g. compounds x targets in a biochemical screen; register a profile with register_fit_profile and pass the result to apply_fit.

Usage

convert_dt_to_se(dt, row_cols, col_cols, assay_name)

Arguments

dt

data.table with one row per measurement.

row_cols

character vector of columns identifying an SE row.

col_cols

character vector of columns identifying an SE column.

assay_name

string; name of the assay.

Value

SummarizedExperiment with row_cols in rowData, col_cols in colData and the remaining columns nested in assay assay_name.

Details

Several row_cols or col_cols are pasted with "_" into the row or column names. Combinations absent from dt come out as empty cells.

Examples

dt <- data.table::CJ(compound = c("cA", "cB"), target = "t1", Concentration = c(0.1, 1))
dt$Activity <- c(90, 40, 80, 20)
convert_dt_to_se(dt, "compound", "target", "Biochemical")
#> class: SummarizedExperiment 
#> dim: 2 1 
#> metadata(0):
#> assays(1): Biochemical
#> rownames(2): cA cB
#> rowData names(1): compound
#> colnames(1): t1
#> colData names(1): target