This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Error while converting Gene ID to Ensembl IDs

I have a DEGs data frame with Gene IDs. Pic for reference below

DEGs List

I am trying to convert the Gene_IDs into Ensembl IDs. I have tried the following methods

library("AnnotationDbi")
library("org.Hs.eg.db")
res3$ensid = mapIds(org.Hs.eg.db,
                      keys=res3$Gene_ID,
                    column="ENSEMBL",
                    keytype = "SYMBOL",
                    multiVals = "first")

The above code converted most of the gene IDs but gave NA values for a couple of them. Can someone please shed some light on this as I can't understand why?

Also tried with biomaRt package

library("biomaRt")
listMarts()
ensembl <- useMart("ensembl")
datasets <- listDatasets(ensembl)
ensembl = useDataset("hsapiens_gene_ensembl", mart = ensembl)
options(max.print = 1000000)
res3$ensid <- getBM(attributes = c('external_gene_name','ensembl_gene_id'), filters = 'external_gene_name',
               values = res3$Gene_ID, mart = ensembl, uniqueRows = FALSE)

but giving the following error

Error in `$<-.data.frame`(`*tmp*`, ensid, value = list(external_gene_name = c("KRT23",  : 
  replacement has 16202 rows, data has 17281

which shows there are more number of rows with Gene IDs in the data. can someone please guide me? Thank you!

ensembl r biomart rna-seq
• 3,640 views • link
updated by jv jv • written by kai_bio • kai_bio

It should be expected that there will be incomplete mappings between different annotation systems. Each [annotation] system has different rules about what to annotate, i.e., what to include. In particular, when dealing with gene symbols, one should expect a difficulty. You could look at some of those genes that are not mapping, and then investigate further. They may be using some 'alias' that is not the official gene symbol, or they may relate to some obscure predicted gene or non-coding RNA that is not even validated.

• 0 views • link
written by Kevin Blighe Kevin Blighe

Thanks for the explanation

• 0 views • link
written by kai_bio • kai_bio

As Kevin said there won't always be a perfect mapping between different ID types. One option is to join the two datasets and fill in NA for those that are missing.

library("dplyr")

ids <- getBM(
  attributes = c('external_gene_name','ensembl_gene_id'), filters = 'external_gene_name',
  values = res3$Gene_ID, mart = ensembl, uniqueRows = FALSE)

res3 <- left_join(res3, ids, by=c("Gene_ID"="external_gene_name"))
• 0 views • link
written by rpolicastro rpolicastro

Thank you! It's working and I was able to join the Ensembl Ids to my available gene ids.

• 0 views • link
written by kai_bio • kai_bio

As others stated, not all gene symbols will have an id in another database. Sometimes this is due to changes in gene annotations, for instance if you are using gene symbols from an older genome assembly versus what is in org.Hs.eg.db which is updated regularly (an issue I have encountered before).

One additional option for pulling more info for your genes, like possible alias symbols, is to query the HGNC Rest API https://www.genenames.org/help/rest/

• 0 views • link
written by jv jv

0 answers

No answers yet.

Log in to answer this question.

More posts like this