InferCNV Algorithm

Purpose

The inferCNV tool is used to infer copy number variations (CNVs) from spatial transcriptomics data, helping to reveal genomic variation characteristics in different tissue regions. This tool can also perform CNV inference on scRNA-Seq data.

Usage

Use the h5ad after cell annotation, prepare the rds file:

SDAS dataProcess h5ad2rds -i st.h5ad -o outdir

Run inferCNV

SDAS infercnv -i st.rds --h5ad st.h5ad --bin_size 50 --slice_key batch \
-o outdir --label_key anno_cell2location --species human \
--ref_group_names B,T --min_counts_per_cell 200

Input Parameter Description

Parameter

Required

Default Value

Description

--run_mode

stRNA

Choose spatial transcriptomics (stRNA) or single-cell (scRNA) mode

-i / --input

Yes

Input Stereo-seq or scRNA data in Seurat rds format, must contain the raw expression matrix

-o / --output

Yes

Output directory containing all results

--h5ad

Yes

h5ad format sample.h5ad, used for spatial heatmap (not required for scRNA mode)

--bin_size

Yes

Bin size, controls spatial heatmap spot size (e.g., 20, 50, 100), not required for scRNA mode

--label_key

Yes

The annotations field, containing cell type or grouping information in rds metadata

--ref_group_names

Reference group names, specify normal cell/sample groups, by default uses all cells (not recommended)

--gene_order_file

Data file containing the positions of each gene along each chromosome in the genome, tab-delimited

--cluster_heatmap

False

Whether to cluster CNV heatmap, True or False

--species

human

Gene location for the specified species (options: human or mouse, default: human). This parameter is invalid when --gene_order_file is set

--slice_key

sampleID

Column name in h5ad.obs indicating slice number for multiple slices

--gene_symbol_key

real_gene_name

The key of gene symbol in meta.data, If set to '_index', treat rownames in rds file as gene symbol

--assay

Assay name in rds used for CNV calculation, if not set, use default assay

--cutoff

0.02

Cut-off for the min average read counts per gene among reference cells/bins

--min_counts_per_cell

100

Mimimun counts allowed per cell/bins

Output Results

Output File

Description

<input_name>_run.final.infercnv_obj.rds

rds object containing the CNV matrix for all genes and spots

<input_name>_CNV_score.csv

CNV score for each spot

<input_name>_CNV_ref.png/pdf

CNV expression heatmap for reference cells (not output if ref_group_names is None)

<input_name>_CNV_obs.png/pdf

CNV expression heatmap for observed cells

<input_name>_CNV_score.png/pdf

Spatial heatmap of CNV score (one per slice for multiple slices, not output in scRNA mode)

CNV Expression Heatmap for Reference Cells: <input_name>_CNV_ref.png/pdf X-axis: spot, Y-axis: gene, color: CNV intensity

CNV Expression Heatmap for Observed Cells: <input_name>_CNV_obs.png/pdf X-axis: spot, Y-axis: gene, color: CNV intensity
- --cluster_heatmap False: Spots are not clustered