RunBanksy()

Spatial transcriptomics clustering incorporating neighborhood gene expression context

RunBanksy() computes a BANKSY matrix that augments each cell’s expression with the average expression of its spatial neighbors. The resulting matrix enables clustering that captures both cell type identity and spatial tissue domains.

Syntax

RunBanksy(
  object,
  lambda,
  assay = "RNA",
  slot = "data",
  use_agf = FALSE,
  dimx = NULL,
  dimy = NULL,
  dimz = NULL,
  ndim = 2,
  features = "variable",
  group = NULL,
  split.scale = TRUE,
  k_geom = 15,
  n = 2,
  sigma = 1.5,
  alpha = 0.05,
  k_spatial = 10,
  spatial_mode = "kNN_median",
  assay_name = "BANKSY",
  M = NULL,
  verbose = TRUE
)

Parameters

object · Seurat · required

A Seurat object with spatial coordinate metadata.

lambda · numeric · required

Spatial weight parameter. 0 uses only cell-intrinsic expression (equivalent to standard clustering). 1 uses only neighborhood expression. Values around 0.2 typically give good cell-type results; 0.8 is better for tissue domain segmentation.

assay · character · default: "RNA"

Assay in the Seurat object to use.

slot · character · default: "data"

Slot within the assay to use.

use_agf · logical · default: "FALSE"

Whether to use the Azimuthal Gabor Filter (AGF) for higher-order neighborhood features.

dimx · character · default: "NULL"

Column name in cell metadata containing the spatial x-coordinate.

dimy · character · default: "NULL"

Column name in cell metadata containing the spatial y-coordinate.

dimz · character · default: "NULL"

Column name in cell metadata containing the spatial z-coordinate (3D data only).

ndim · integer · default: "2"

Number of spatial dimensions to use (2 or 3).

features · character · default: "variable"

Features to compute. One of "all", "variable", or a character vector of feature names.

group · character · default: "NULL"

Column name of a grouping variable in metadata. Used to compute neighbors within groups.

split.scale · logical · default: "TRUE"

Whether to scale expression separately by group.

k_geom · numeric · default: "15"

Number of neighbors for the kNN spatial graph (used with kNN-based spatial_mode).

n · numeric · default: "2"

Exponent of radius for kNN_rn spatial mode.

sigma · numeric · default: "1.5"

Standard deviation of the Gaussian kernel for rNN_gauss mode.

alpha · numeric · default: "0.05"

Determines the radius used in rNN_gauss mode.

k_spatial · numeric · default: "10"

Number of neighbors for radial nearest neighbor modes.

spatial_mode · character · default: "kNN_median"

Kernel for spatial neighborhood computation. Options:

• kNN_median — k-nearest neighbors with median-scaled Gaussian kernel
• kNN_r — k-nearest neighbors with 1/r kernel
• kNN_rn — k-nearest neighbors with 1/r^n kernel
• kNN_rank — k-nearest neighbors with rank Gaussian kernel
• kNN_unif — k-nearest neighbors with uniform kernel
• rNN_gauss — radial nearest neighbors with Gaussian kernel

assay_name · character · default: "BANKSY"

Name for the new BANKSY assay stored in the Seurat object.

M · numeric · default: "NULL"

Highest azimuthal harmonic for AGF computation. Advanced usage.

verbose · logical · default: "TRUE"

Print progress messages.

Returns

A Seurat object with a new assay (default name BANKSY) containing the BANKSY-augmented expression matrix. Downstream analysis (PCA, UMAP, clustering) should be run on this assay.

Examples

library(SeuratWrappers)

# Run BANKSY with spatial coordinates in metadata
object <- RunBanksy(
  object = seurat_obj,
  lambda = 0.2,
  assay = "RNA",
  slot = "data",
  dimx = "x",
  dimy = "y",
  features = "variable",
  spatial_mode = "kNN_median",
  k_geom = 15
)

# Downstream clustering on BANKSY assay
DefaultAssay(object) <- "BANKSY"
object <- ScaleData(object)
object <- RunPCA(object)
object <- FindNeighbors(object)
object <- FindClusters(object, resolution = 0.5)

See Also

• BANKSY method guide