Mapping Entrez gene ids to Ensembl ids

Question

[Natasha]

I am mapping Entrez gene ids to ensembl ids using biomaRt.

library("biomaRt")                                                                                                                    
library("readxl")                                                                                                                     
tbl <- read_excel(input_excel)                                                                                                   
tbl <- tbl["EntrezID"]                                                                                                                
print(tbl, n=Inf)                                                                                                                     
listMarts()                                                                                                                           
ensembl <- useMart("ensembl",dataset="hsapiens_gene_ensembl")                                                                         
filters = listFilters(ensembl)                                                                                                        
entrezgene = tbl                                                                                                                                                                                                                                   
genes <- getBM(                                                                                                                               
filters="entrezgene_id",                                                                                                              
attributes=c("ensembl_gene_id","entrezgene_id"),                                                                                      
values=entrezgene,                                                                                                                    
mart=ensembl)                                                                                                                 
print(genes)

The input is read from the excel here from this study.

I get the following error message

Error in `[[<-.data.frame`(`*tmp*`, vIdx, value = c("26155", "9636", "375790",  :                                                       replacement has 494 rows, data has 6909                                                                                            
 Calls: getBM ... .generateFilterXML -> .splitValues -> [[<- -> [[<-.data.frame                                                        
Execution halted                                                                                                                      
Error in `[[<-.data.frame`(`*tmp*`, vIdx, value = c("26155", "9636", "375790",  :                                                       replacement has 494 rows, data has 6909                                                                                             
Calls: getBM ... .generateFilterXML -> .splitValues -> [[<- -> [[<-.data.frame                                                        
Execution halted

Does 494 rows, data has 6909 this mean only ensembl ids of 494 entrez genes(out of 6909) were mapped? Could someone look into this? I'm not sure how to resolve this error.

[MatthewP]

Maybe try convert entrez_id from character to nemuric?

[Natasha]

Thank you. I added tbl <- mapply(tbl, FUN=as.numeric) before listMart(). The query proceeds without any error. However, I obtain 7720 entries, whereas my input has 6909 entries.

[MatthewP]

This is normal, many ENTREZ ID map to multiple ENSEMBL ID.

[Haci]

Can you post the actual function that gives the error? Where is your getBM() call?

[Natasha]

Sorry, there was a formatting error. Now it's displayed

Answer 1

[Mike Smith]

The fundamental issue here is that your entrezgene object is a tibble with one column, where as the getBM() function expects to be passed a vector. The internal code doesn't handle this very well and throws the error you saw.

When you run tbl <- mapply(tbl, FUN=as.numeric) you get back a matrix with a single column, which still isn't what the function expects, but it seems to handle it better. However it is not required to provide the IDs as numbers (in fact they will be converted back to characters inside getBM() anyway).

The better solution is to just select the appropriate column from your tbl an return it as a vector, which you could do via: entrezgene <- tbl$EntrezID

I will update biomaRt to check it's being given the correct type of input, and provide some guidance if that's not the case.