Skip to contents

Automated ROI detection using density-based spatial clustering

Source: R/NicheDetect.R
NicheDetect_STS.Rd

This function automatically detects regions of interest (ROIs) using density-based clustering (DBSCAN) on spatial transcriptomics data. It includes preprocessing steps like density filtering and iterative spot updating.

Usage

NicheDetect_STS(
  STID_obj = NULL,
  meta_key = NULL,
  loop_id = "LoopAllSamp",
  pos_colnm = NULL,
  neg_value = "neg",
  spatial_scale_method = "region",
  region_detect_method = "convex",
  concavity = 2,
  update_spots = TRUE,
  density_thres = 0.9,
  ROI_size = 10,
  minPts = NULL,
  k_kNNdist = NULL,
  description = NULL,
  grp_nm = NULL,
  dir_nm = "M2_NicheDetect_STS"
)

Arguments

STID_obj

An STID object containing spatial transcriptomics data

meta_key

Character, metadata key containing positive spot information

loop_id

Character, sample grouping identifier (default: "LoopAllSamp")

pos_colnm

Character, column name containing positive spot labels

neg_value

Character, value indicating negative/non-ROI spots (default: "neg")

spatial_scale_method

Character, detection level - "region" or "spot" (default: "region")

region_detect_method

Character, hull method - "convex" or "concave" (default: "convex")

concavity

Numeric, concavity parameter for concave hull (default: 2)

update_spots

Logical, whether to iteratively update positive spots (default: TRUE)

density_thres

Numeric, density threshold for filtering (0-1, default: 0.9)

ROI_size

Integer, minimum number of spots to form an ROI (default: 10)

minPts

Integer, minimum points parameter for DBSCAN (default: NULL, automatically determined based on data format)

k_kNNdist

Integer, k value for kNN distance calculation (default: NULL, uses minPts if not specified)

description

Character, description of the analysis (default: NULL)

grp_nm

Character, group name for output organization (default: NULL, uses timestamp)

dir_nm

Character, directory name for output (default: "M2_NicheDetect_STS")

Value

Returns the modified STID object with added metadata containing ROI labels, distances to ROI centers, edge information, and region classifications

Examples

if (FALSE) { # \dontrun{
# Automatically detect ROIs based on positive spots
STID_obj <- NicheDetect_STS(
  STID_obj = ist_object,
  meta_key = "M1_SpotDetect_Gene_20240101",
  pos_colnm = "Label_geneA",
  density_thres = 0.9,
  ROI_size = 10
)
} # }