Persist a data.table into a BumpyMatrix assay — persist_fit

Performs an idempotent upsert (merge mode) keyed by upsert_key_cols, or a full overwrite (replace mode).

Usage

persist_fit_assay(se, new_dt, merge, assay_name, row, col, upsert_key_cols)

Arguments

se: SummarizedExperiment
new_dt: data.table including "row" and "column" index fields
merge: "merge" or "replace"
assay_name: assay name to write into
row: row index column name in new_dt
col: column index column name in new_dt
upsert_key_cols: character vector of column(s) forming the upsert key

Value

updated SummarizedExperiment

Examples

mae <- get_synthetic_data("finalMAE_small.qs2")
se <- mae[["single-agent"]]
dt <- data.table::data.table(
  row = rownames(se)[1], column = colnames(se)[1],
  fit_source = "demo", x_mean = 0.5
)
se <- persist_fit_assay(se, dt, "merge", "custom_mean", "row", "column", "fit_source")
SummarizedExperiment::assayNames(se)
#> [1] "RawTreated"  "Controls"    "Normalized"  "Averaged"    "Metrics"    
#> [6] "custom_mean"