Biomart query returns NA when searching for entrez_id, while manual search works

Question

[eggrandio]

Hi,

I am using biomart to convert gene IDs into entrezid accessions. I am working with arabidopsis genes, and sometimes, the query will not return an entrezid. However, when I look up that gene in the ensembl webpage, I am able to find the correspondence. Maybe I am using the wrong Mart/Dataset?

Here is the code I am using for reference and the results I obtain:

library(biomaRt)
ensembl = useMart("plants_mart",host="plants.ensembl.org")
ensembl = useDataset("athaliana_eg_gene",mart=ensembl)
genes = c("AT2G14610","AT4G23700","AT3G26830","AT3G15950","AT3G54830","AT5G24105")
query = getBM(attributes=c("ensembl_gene_id",
                       "entrezgene_id",
                       "refseq_dna",
                       "entrezgene_accession"),
          filters=("ensembl_gene_id"),
          values=genes,mart=ensembl)

> query
   ensembl_gene_id entrezgene_id     refseq_dna entrezgene_accession
1        AT2G14610        815949    NM_127025.3               815949
2        AT3G15950        820839    NM_112465.4               820839
3        AT3G15950        820839 NM_001035631.2               820839
4        AT3G15950        820839 NM_001338192.1               820839
5        AT3G15950        820839 NM_001338191.1               820839
6        AT3G15950        820839 NM_001338193.1               820839
7        AT3G26830        822298    NM_113595.4               822298
8        AT3G54830            NA                                  NA
9        AT4G23700        828470 NM_001341626.1               828470
10       AT4G23700        828470    NM_118501.5               828470
11       AT5G24105       2745995    NM_203099.2              2745995

In this case, AT3G54830 does not show any entrezgene_id or refseq_dna. However, when I manually search for it at the plant.ensembl.org or NCBI webpages, I can find it:

https://plants.ensembl.org/Arabidopsis_thaliana/Transcript/Summary?db=core;g=AT3G54830;r=3:20311901-20315887;t=AT3G54830.1

Or when I search for it in the NCBI webpage:

https://www.ncbi.nlm.nih.gov/gene/824648

Any help would be appreciated!

Thanks!

[Kevin Blighe]

biomaRt just interacts with the internal servers at Ensembl, so, just acts as an interface to whatever is held at Ensembl. This gene just appears to not yet have a NM ID, but it is listed under the refseq_peptide attribute.

[eggrandio]

But when I look it up in the NCBI website, it returns entrez_id 824648, and once there, I can find the NM IDs: NM_001339700.1 and NM_115340.3 corresponding to those two refseq_peptide entries.

I guess it could be because the database at ensembl is not updated? I find it weird, as it looks like the original sequence of the transcript was uploaded in 2016.

Is there any way of querying NCBI database with biomart?

[Kevin Blighe]

I see what you mean. It can be that Ensembl's databases have not yet updated - not sure how it works internally. In terms of automated annotation, though, there are basically 2 main ways:

• biomaRt
• ord.db packages

Each has pros and cons.

In your case, it seems better to use org.db (see answer below)

Answer 1

[Kevin Blighe]

Answer:

library(org.At.tair.db)

genes <- c("AT2G14610","AT4G23700","AT3G26830",
  "AT3G15950","AT3G54830","AT5G24105")

keytypes(org.At.tair.db)

 [1] "ARACYC"       "ARACYCENZYME" "ENTREZID"     "ENZYME"       "EVIDENCE"    
 [6] "EVIDENCEALL"  "GENENAME"     "GO"           "GOALL"        "ONTOLOGY"    
[11] "ONTOLOGYALL"  "PATH"         "PMID"         "REFSEQ"       "SYMBOL"      
[16] "TAIR"        


mapIds(org.At.tair.db, keys = genes,
  column = c('ENTREZID'), keytype = 'TAIR')
'select()' returned 1:1 mapping between keys and columns
AT2G14610 AT4G23700 AT3G26830 AT3G15950 AT3G54830 AT5G24105 
 "815949"  "828470"  "822298"  "820839"  "824648" "2745995" 



select(org.At.tair.db, keys = genes,
  column = c('ENTREZID', 'SYMBOL', 'REFSEQ'), keytype = 'TAIR')

'select()' returned 1:many mapping between keys and columns
        TAIR ENTREZID   SYMBOL       REFSEQ
1  AT2G14610   815949    ATPR1    NM_127025
2  AT2G14610   815949    ATPR1    NP_179068
3  AT2G14610   815949       PR    NM_127025
4  AT2G14610   815949       PR    NP_179068
5  AT2G14610   815949      PR1    NM_127025
6  AT2G14610   815949      PR1    NP_179068
7  AT4G23700   828470  ATCHX17 NM_001341626
8  AT4G23700   828470  ATCHX17    NM_118501
9  AT4G23700   828470  ATCHX17 NP_001328705
10 AT4G23700   828470  ATCHX17    NP_194101
11 AT4G23700   828470    CHX17 NM_001341626
12 AT4G23700   828470    CHX17    NM_118501
13 AT4G23700   828470    CHX17 NP_001328705
14 AT4G23700   828470    CHX17    NP_194101
15 AT3G26830   822298 CYP71B15    NM_113595
16 AT3G26830   822298 CYP71B15    NP_189318
17 AT3G26830   822298     PAD3    NM_113595
18 AT3G26830   822298     PAD3    NP_189318
19 AT3G15950   820839     NAI2 NM_001035631
20 AT3G15950   820839     NAI2 NM_001338191
21 AT3G15950   820839     NAI2 NM_001338192
22 AT3G15950   820839     NAI2 NM_001338193
23 AT3G15950   820839     NAI2    NM_112465
24 AT3G15950   820839     NAI2 NP_001030708
25 AT3G15950   820839     NAI2 NP_001326807
26 AT3G15950   820839     NAI2 NP_001326808
27 AT3G15950   820839     NAI2 NP_001326809
28 AT3G15950   820839     NAI2    NP_188216
29 AT3G54830   824648     <NA> NM_001339700
30 AT3G54830   824648     <NA>    NM_115340
31 AT3G54830   824648     <NA> NP_001326240
32 AT3G54830   824648     <NA>    NP_191043
33 AT5G24105  2745995    AGP41    NM_203099
34 AT5G24105  2745995    AGP41    NP_974828

[eggrandio]

Thanks for your reply!

A while ago I tested both ways (for different queries, related with GOs) and it seemed that ensembl through biomaRt was more complete, but in this case it looks like the contrary.

I guess it depends on the query... I wish there was a unified database.

I might have to write an script that searches both ways and unifies the results.