Tumor Microenvironment (TME) Deconvolution Pipeline

Executes an integrated TME analysis on a gene expression matrix: performs immune/stromal cell deconvolution using multiple algorithms, computes signature scores, and aggregates results. Designed for exploratory immunogenomic profiling.

Usage

iobr_deconvo_pipeline(
  eset,
  project,
  array,
  tumor_type,
  path = "1-TME",
  permutation = 1000
)

Arguments

eset

Numeric matrix. Gene expression (TPM/log scale) with genes in rows.

project

Character. Project name (used in output naming).

array

Logical. Whether data originated from an array platform. Affects deconvolution choices.

tumor_type

Character. Tumor type code (e.g., "stad") used by certain methods.

path

Character. Output directory. Default is "1-TME".

permutation

Integer. Number of permutations for CIBERSORT (and similar). Default is 1000.

Value

Data frame integrating cell fractions and signature scores (also writes intermediate outputs to disk).

Author

Dongqiang Zeng

Examples

eset_stad <- load_data("eset_stad")
anno_grch38 <- load_data("anno_grch38")
eset <- anno_eset(eset = eset_stad, annotation = anno_grch38, probe = "id")
#> ℹ Row number of original eset: 60483
#> ✔ 100% of probes in expression set were annotated
#> ℹ Found 2293 duplicate symbols, using "mean" method
#> ℹ Row number after filtering duplicated gene symbol: 50181
# \donttest{
res <- iobr_deconvo_pipeline(
  eset = eset, project = "STAD",
  array = FALSE, tumor_type = "stad",
  path = tempdir(), permutation = 10
)
#> ✔ Applied log2 transformation
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running CIBERSORT
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running EPIC deconvolution
#> Warning: The optimization didn't fully converge for some samples:
#> TCGA-BR-6455; TCGA-FP-7916
#>  - check fit.gof for the convergeCode and convergeMessage
#> Warning: mRNA_cell value unknown for some cell types: CAFs, Endothelial - using the default value of 0.4 for these but this might bias the true cell proportions from all cell types.
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running MCP-counter deconvolution
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running xCell deconvolution
#> ℹ Number of genes: 10783
#> ℹ GSVA version 2.4.8
#> ℹ Searching for rows with constant values
#> ℹ Calculating GSVA ranks
#> ℹ kcdf='auto' (default)
#> ℹ GSVA dense (classical) algorithm
#> ℹ Row-wise ECDF estimation with Gaussian kernels
#> ℹ Calculating row ECDFs
#> ℹ Calculating column ranks
#> ℹ GSVA dense (classical) algorithm
#> ℹ Calculating GSVA scores
#> ✔ Calculations finished
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running ESTIMATE
#> [1] "Merged dataset includes 10148 genes (264 mismatched)."
#> [1] "1 gene set: StromalSignature  overlap= 138"
#> [1] "2 gene set: ImmuneSignature  overlap= 140"
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running TIMER deconvolution
#> ℹ Enter batch mode
#> ℹ Loading immune gene expression
#> ℹ Outlier genes: ACTB ACTG1 CD74 COL1A1 EEF1A1 ERBB2 FLNA IGHG1 IGKC MT-CO1 MT-CO2 MT-ND4 MT-RNR2 MYH11
#> ℹ Removing batch effects for stad
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running quanTIseq deconvolution
#> ℹ Running quanTIseq deconvolution module
#> ℹ Gene expression normalization and re-annotation (arrays: FALSE)
#> ℹ Removing 17 noisy genes
#> ℹ Removing 15 genes with high expression in tumors
#> ℹ Signature genes found in data set: 138/138 (100%)
#> ℹ Mixture deconvolution (method: lsei)
#> ✔ Deconvolution successful!
#> Warning: Data values appear small (< 50).
#> ℹ Input should be in TPM/FPKM scale, not log-transformed
#> ℹ Running IPS calculation
#> [1] ">>>>> TME cell deconvolution was completed: STAD"
#> ℹ Calculating signature scores using PCA, z-score, and ssGSEA methods
#> ℹ Log2 transformation not necessary (data appears to already be log-scaled)
#> ℹ Step 1/3: PCA method
#> ℹ Step 2/3: z-score method
#> ℹ Step 3/3: ssGSEA method
#> ℹ GSVA version 2.4.8
#> ℹ Searching for rows with constant values
#> ℹ Calculating ssGSEA scores for 280 gene sets
#> ℹ Calculating ranks
#> ℹ Calculating rank weights
#> ℹ Normalizing ssGSEA scores
#> ✔ Calculations finished
#> [1] ">>>>> Signature esitmation was completed: STAD"
#> ℹ Calculating signature scores using ssGSEA method
#> ℹ Log2 transformation not necessary (data appears to already be log-scaled)
#> ℹ Calculating scores for 12059 signature(s)
#> ℹ GSVA version 2.4.8
#> ℹ Searching for rows with constant values
#> ℹ Calculating ssGSEA scores for 12059 gene sets
#> ℹ Calculating ranks
#> ℹ Calculating rank weights
#> ℹ Normalizing ssGSEA scores
#> ✔ Calculations finished
#> [1] ">>>>> HALLMARK GO KEGG REACTOME esitmation was completed: STAD"
# }