Atera's per-transcript codeword_identity (which specific barcode/
codeword called each detection) is collapsed away by
LoadAteraMolecules, which reports one combined position per
gene. A gene can map to more than one distinct codeword (more than one
probe/barcode design targeting the same gene); this instead returns one
"<gene>_codeword_<id>" pseudo-feature per codeword, which is useful
for a common QC question: whether different codewords for the same gene
cluster differently in space (which would suggest a probe-specific
artifact rather than true expression).
LoadAteraCodewords(object, data.dir, genes, fov = NULL)object, with one "<gene>_codeword_<id>" molecule
entry per codeword mapping to each requested gene, merged into
object[[fov]]'s existing molecule layer (regular genes loaded via
LoadAteraMolecules are preserved, not overwritten; genes
already broken down by a previous call are not re-fetched)
A Seurat object created by LoadAtera
The same data.dir passed to the original
LoadAtera call (the Atera output bundle directory)
Character vector of one or more gene names to break down by codeword
Name of the FOV to attach the resulting molecules to; defaults to the object's default FOV