Unable to access ensembl biomart archives via R

Question

[johntimcham]

EDIT: I tried updating biomaRt, which replaced it with version 2.38.0 (see my response to genomax). After restarting RStudio the error has vanished. Perhaps there was an issue with some dependency.

I'm trying to access biomart version 93 (human) because this happens to be the version that was used to build the reference GTF I've been using. This previous thread had the answer: BiomaRt: archive version of ensembl not working

Unfortunately the system seems to be broken. The listEnsemblArchives() function returns the error:

Error in getNodeSet(html, path = "//div[@class='plain-box float-right archive-box']")[[1]] : 
  subscript out of bounds
In addition: Warning message:
XML content does not seem to be XML: 'https://www.ensembl.org/info/website/archives/index.html?redirect=no'

Similarly, the following commands returns the identical error:

hg93 <- useEnsembl(biomart = "ensembl", dataset = "hsapiens_gene_ensembl",
     version = "93")
hg93 <- useEnsembl(biomart = "ENSEMBL_MART_ENSEMBL", dataset = "hsapiens_gene_ensembl", 
    version = "93")
hg93 <- useMart("ensembl",host = "http://jul2018.archive.ensembl.org",
    dataset = "hsapiens_gene_ensembl")

The current version works fine. I normally use this format:

useMart("ENSEMBL_MART_ENSEMBL","hsapiens_gene_ensembl")

I'm not sure if there is any difference between useMart() and useEnsembl()

Thank you!

[xiaohk2016]

I experienced the same error today. And then I tried to open the wedsite:https://www.ensembl.org/info/website/archives/index.html,but it showed "The Ensembl web service you requested is temporarily unavailable". It might be the reason.

[GenoMax]

mart <- useEnsembl(biomart = "ensembl", dataset = "hsapiens_gene_ensembl", version = "93")
listEnsemblArchives()
             name     date                                url version current_release
1  Ensembl GRCh37 Feb 2014          http://grch37.ensembl.org  GRCh37                
2      Ensembl 97 Jul 2019 http://jul2019.archive.ensembl.org      97               *
3      Ensembl 96 Apr 2019 http://apr2019.archive.ensembl.org      96                
4      Ensembl 95 Jan 2019 http://jan2019.archive.ensembl.org      95                
5      Ensembl 94 Oct 2018 http://oct2018.archive.ensembl.org      94                
6      Ensembl 93 Jul 2018 http://jul2018.archive.ensembl.org      93                
7      Ensembl 92 Apr 2018 http://apr2018.archive.ensembl.org      92                
8      Ensembl 91 Dec 2017 http://dec2017.archive.ensembl.org      91                
9      Ensembl 90 Aug 2017 http://aug2017.archive.ensembl.org      90                
10     Ensembl 89 May 2017 http://may2017.archive.ensembl.org      89                
11     Ensembl 88 Mar 2017 http://mar2017.archive.ensembl.org      88                
12     Ensembl 87 Dec 2016 http://dec2016.archive.ensembl.org      87                
13     Ensembl 86 Oct 2016 http://oct2016.archive.ensembl.org      86                
14     Ensembl 85 Jul 2016 http://jul2016.archive.ensembl.org      85

Works for me without an error.

As for your other question:

The useEnsembl function enables one to connect to a specified BioMart database and dataset hosted by Ensembl without having to specify the Ensembl URL.

[johntimcham]

What version of R and biomaRt are you using?

I am on:

R version 3.5.1 (2018-07-02)

biomaRt_2.41.4

edit: I tried updating my biomart, which resulted in it being replace by biomaRt_2.38.0. I am on bioconductor version 3.8 because my university's compute cluster hasn't installed R 3.6 yet.

I'm not sure where 2.41.4 came from given that the latest version listed on https://bioconductor.org/packages/release/bioc/html/biomaRt.html is version 2.40.3

Anyway the error remains

Thanks

Answer 1

[Mike Smith]

Are you still experiencing this error? I get the same correct response as Genomax when I test your example code.

If it's now working for you then I suspect it was a temporary outage of the Ensembl site, and biomaRt didn't handle it very gracefully. If not the please post the output of sessionInfo() here and I'll try to recreate your setup to identify the problem.

The 2.41.x versions are from the developmental branch, which can be installed from Bioconductor or Github, but shouldn't be considered stable. However I don't think I've made an recent changes that should introduce this problem.

[GenoMax]

Mike see this addendum by OP at top of the post. Looks like the problem was resolved after an update:

I tried updating biomaRt, which replaced it with version 2.38.0 (see my response to genomax). After restarting RStudio the error has vanished. Perhaps there was an issue with some dependency.