What does the "filters" argument in "getBM" do?

Question

In the biomaRt package, there is a function getBM which among a few things is useful for mapping between different gene representations. Currently, I'm using it as follows, to map Ensembl transcript IDs to gene names and Ensembl gene IDs.

ensemble2gene <- getBM(attributes=c("ensembl_transcript_id", 
                                    "external_gene_name",
                                    "ensembl_gene_id"),
                       values = as.list(transcripts), 
                       mart = mart)

How is the following code different:

ensemble2gene <- getBM(attributes=c("ensembl_transcript_id", 
                                    "external_gene_name",
                                    "ensembl_gene_id"),
                       filters = "ensembl_transcript_id",
                       values = as.list(transcripts), 
                       mart = mart)

It's unclear to me from the description of the filters argument here. But just based on results, I get less results when I run the second one.

Answer 1

Hi An Ignorant Wanderer,

Filters allow you to narrow down the dataset to query a specific subset of features within your dataset. A description and tutorial of the filtering steps within BioMart can be found here: https://www.ensembl.org/info/data/biomart/how_to_use_biomart.html

If you remove filters, you will query the full dataset (all human genes, in the example above). However, this is not recommended for BioMart as large, genome-wide queries will overload the BioMart servers. For queries of this size, we suggest using the Ensembl REST API or FTP site.