Hi Nuorain,

Thank you for sharing the screenshot of your environment.

Using the in-house dataset dataBRCA for tabDF, I noticed the row names are Entrez Gene IDs (used by NCBI) and gene symbols separated by a vertical bar, and are not Ensembl IDs:

>  head(dataBRCA)
TCGA-A7-A13D-01A-13R-A12P-07 TCGA-BH-A0DK-11A-13R-A089-07 TCGA-BH-A1FC-11A-32R-A13Q-07 TCGA-E9-A1RH-11A-34R-A169-07
        ?|100133144                        20.47                        54.59                         7.84                         6.05
        ?|100134869                         9.53                         8.41                         9.16                         8.95
        ?|10357                           258.79                       241.38                       108.00                       229.01

It's possible but not ideal nor free of potential issues to map these Entrez Gene IDs to Ensembl IDs. Here is a first-order approach with print statements and comments, if helpful.

### Load dependencies. ###
if (!require("BiocManager", quietly = TRUE))
  install.packages("BiocManager")
BiocManager::install("TCGAbiolinks")
BiocManager::install("EDASeq")


#Load. 
library(TCGAbiolinks)
library(EDASeq)
library(biomaRt)
library(dplyr)

### Prepare dataBRCA. ###
#Retain integer Entrez Gene ID part of your query data.
entrez_gene_ids <- rownames(dataBRCA)
entrez_gene_ids_clean <- as.character(sub(".*\\|", "", entrez_gene_ids))
#Retrieve Ensembl gene IDs.
ensembl_human <- useMart("ensembl", dataset = "hsapiens_gene_ensembl")
gene_mapping <- getBM(
  attributes = c("entrezgene_id", "ensembl_gene_id", "hgnc_symbol"),
  filters = "entrezgene_id",
  values = entrez_gene_ids_clean,
  mart = mart
)
gene_mapping$entrezgene_id <- as.character(gene_mapping$entrezgene_id)
gene_mapping$hgnc_symbol <- as.character(gene_mapping$hgnc_symbol)

#Subset.
entrez_gene_ids_clean <- strsplit(entrez_gene_ids, "\\|")
before_pipe <- sapply(entrez_gene_ids_clean, `[`, 1)  # Part before "|"
after_pipe <- sapply(entrez_gene_ids_clean, `[`, 2)   # Part after "|"
#Method 1: excludes entries with ? for gene symbol.
# subset_ME <- gene_mapping[gene_mapping$entrezgene_id %in% after_pipe
#                 & gene_mapping$hgnc_symbol %in% before_pipe, ]
#Method 2: includes entries with ? for gene symbol, and ensemble ID for 
#genes with identical entrezgeneid and gene symbol.
overlap_gene_symbol <- intersect(gene_mapping$hgnc_symbol, before_pipe)
length(overlap_gene_symbol) #16865 total overlapping genes by symbol between dataBRCA and Ensembl.
overlap_geneid <- intersect(gene_mapping$entrezgene_id, after_pipe)
length(overlap_geneid) #19644 total overlapping genes by geneID between dataBRCA and Ensembl.
matches <- rep(FALSE, nrow(gene_mapping))
for (i in seq_along(before_pipe)) {
  matches <- matches | ((before_pipe[i] == "?" | gene_mapping$hgnc_symbol == before_pipe[i]) &
                          gene_mapping$entrezgene_id == after_pipe[i])
}
#> table(matches)
# matches
# FALSE  TRUE 
# 3268 19515
#There are 19515 entries in gene list that matches with our query data.
#Subset matching indices.
#There are 28 duplicated gene symbols, which we cannot
#accurately map to our gene symbol.
subset <- gene_mapping[matches, ]
#Original gene list has 22783 total genes. Query data (dataBRCA) 
#has 20530 total genes. Their overlap differs by type of identifier for genes.
#Overall, there are 19515 matching genes between gene list and query
#based on above logic.

#Order subset of gene list by query data. 
#Find matches. 
subset$match_index <- match(subset$entrezgene_id, after_pipe)
#Order df.
sorted_subset <- subset %>%
  arrange(match_index) %>%
  select(-match_index, -hgnc_symbol)

#For checking.
shared <- match(sorted_subset$entrezgene_id, after_pipe) #19515 original
subset_query <- dataBRCA[shared,]
rownames(subset_query) <- sorted_subset$ensembl_gene_id
#dataBRCA genes now with Ensembl IDs.

#Find sample to test.
#Use index to confirm order in ref gene list.
#matches <- grep("\\b100\\b", after_pipe)

#Normalize. 
norm_data <- TCGAanalyze_Normalization(
  tabDF=subset_query,
  geneInfo=geneInfoHT,
  method="gcContent"
)

> norm_data
                TCGA-A7-A13D-01A-13R-A12P-07 TCGA-BH-A0DK-11A-13R-A089-07 TCGA-BH-A1FC-11A-32R-A13Q-07 TCGA-E9-A1RH-11A-34R-A169-07
ENSG00000000003                         3144                         2951                         4004                         2953
ENSG00000000005                            0                           26                         1082                          571
ENSG00000000419                         2098                         1062                         1381                         1126

Using Ensembl IDs (or the appropriate gene name IDs) as row names for your tabDF object resolves the initial error related to an empty vector.

Maze