mergeTumorIntoX: Estimate a tumor-specific profile and merge it with the...

In SpatialDecon: Deconvolution of mixed cells from spatial and/or bulk gene expression data

Description

Given the input of "tumor-only" AOI's, estimates an collection of tumor-specific expression profiles and merges them with the immune cell expression training matrix. The process:

1. log2/normalized data from tumor-only AOIs is clustered with hclust, and cutree() is used to define clusters.

2. 2. Each cluster's geomean profile is merged into the immune cell profile matrix.

Usage

mergeTumorIntoX(norm, bg, pure_tumor_ids, X, K = 10)

Arguments

norm	matrix of normalized data
bg	matrix of expected background, on the scale of norm.
pure_tumor_ids	Vector identifying columns of norm that are pure tumor. Can be indices, logicals or column names.
X	The training matrix
K	the number of clusters to fit

Value

an updated X matrix with new columns, "tumor.1", "tumor.2", ...

Examples

data(mini_geomx_dataset)
data(safeTME)
mini_geomx_dataset$bg <- derive_GeoMx_background(
  norm = mini_geomx_dataset$normalized,
  probepool = rep(1, nrow(mini_geomx_dataset$normalized)),
  negnames = "NegProbe"
)
safeTME.with.tumor <- mergeTumorIntoX(
  norm = mini_geomx_dataset$norm,
  bg = mini_geomx_dataset$bg,
  pure_tumor_ids = mini_geomx_dataset$annot$AOI.name == "Tumor",
  X = safeTME,
  K = 3
)

SpatialDecon documentation built on Nov. 8, 2020, 6 p.m.