Question

[clusterprofiler] Transformation from ENSEMBL ID to entrez ID or Symbol got reduced number of genes

0

Entering edit mode

tenger929 • 0

@tenger929-18259

Last seen 5.4 years ago

Dear, It is convenient to use clusterprofiler in enrichment analysis. However, there is a problem confusing me in ID transformation.

I have a gene list containing 39570 genes with ensembl ID.

1. After transformation with bitr function from Ensembl to entrez ID, the number of genes is 22142, 44.39% of input gene IDs fail to map.

2. Transformation with bitr function from Ensembl to Symbol, the number of genes is also 22142, however, it contains duplicates. After removing duplicates, the number of genes is 22103.

I don't know why it happened. Any help will be greatly appreciated!

Thanks a lot.

clusterprofiler • 2.8k views

ADD COMMENT • link updated 5.4 years ago by James W. MacDonald 65k • written 5.4 years ago by tenger929 • 0

score 0 · Answer 1 · 2018-12-03

0

Entering edit mode

James W. MacDonald 65k

@james-w-macdonald-5106

Last seen 20 minutes ago

United States

Converting IDs from one annotation service to another is a fraught process. EBI/EMBL and NCBI use different criteria to say what is and isn't a gene, and it is not uncommon to lose a significant proportion of your annotations when converting from one to the other. The best course of action is to stick with one or the other through the entirety of your analysis.

ADD COMMENT • link 5.4 years ago James W. MacDonald 65k

0

Entering edit mode

Thanks for your kindly reply.

My RNAseq (mouse species) was obtained from strand specific RNA library，and we got gene list with Ensembl ID by DESeq2. In enrichment analysis by clusterprofiler developed by Prof. Yu.，Entrez ID is prefered. I have no idear why so large proportion of genes are unmapped during ID transformatiom from Ensembl to Entrez.

ADD REPLY • link 5.4 years ago tenger929 • 0