mapIDs (AnnotationDBi) returns shorter list of IDs than input IDs

Question

[mbk0asis]

Hi, all!

While I was trying to convert Ensembl ID to Entrez ID using 'mapIDs', I noticed number of missing IDs in the result.

18,453 Ensembl IDs were input, but only 18,364 Entrez IDs were returned.

I was expecting the missing IDs would be presented as 'NA' or something like that. Instead, they weren't included in the result at all.

The data I'm trying to convert IDs,

> head(de)
                   p_val avg_logFC pct.1 pct.2 p_val_adj fcsign logP metric
ENSMUSG00000000184     0 1.5573130 0.992 0.016         0      1    0      0
ENSMUSG00000053030     0 1.3555568 1.000 0.005         0      1    0      0
ENSMUSG00000037820     0 1.1589184 0.875 0.009         0      1    0      0
ENSMUSG00000005087     0 0.9234463 0.842 0.002         0      1    0      0
ENSMUSG00000026586     0 0.7931385 0.658 0.000         0      1    0      0
ENSMUSG00000050010     0 0.7784591 0.733 0.000         0      1    0      0

and R code I used,

ranks$entrez = mapIds(org.Mm.eg.db,
                      keys=row.names(de), 
                      column="ENTREZID",
                      keytype="ENSEMBL",
                      multiVals="filter")

What seems to be the problem here?

Please, help me!

Thank you!

Answer 1

[Arup Ghosh]

There is no issue with the annotation, the Ensembl ids with missing Entrez annotation are unique to Ensembl. Check the attached article for an overview of all the annotation databases.

https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-015-1308-8

[mbk0asis]

Thank you for reply! So you're saying is that Ensembl has multiple ID for a Entrez annotation? Am I understanding right? Then how would you handle those situations?

[Arup Ghosh]

No, I meant to say some genomic features only has an annotation only in Ensembl database , not in Entrez.