Search Mailing List Archives

Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] NCBO Annotator WebService

Ray Fergerson ray.fergerson at
Thu Jul 24 14:33:59 PDT 2014


The annotation mechanism is very simple - just string matching. Thus:

- Exact_match = false means to all substring match. No mismatches allowed.

- No singular plural matching and no stemming.

I believe that you can get the behavior you call parent matching by using
the "longest only" match.

If you are creating a package for reuse by others, we ask that you create
an new account for the package (and get its own API key). That way we can
track the usage of your package independent of your personal usage.


-----Original Message-----
From: bioontology-support
[mailto:bioontology-support-bounces at] On Behalf Of
Eugenia Galeota
Sent: Wednesday, July 23, 2014 9:17 AM
To: support at
Subject: [bioontology-support] NCBO Annotator WebService

To whom who may concern,
my name is Eugenia and I'm a post doc at the Center for genomic Science of
the Italian Institute of Technology. I'm trying to annotate GEO metadata
using BRENDA ontology and DISEASE ONTOLOGY and I would like to use the
NCBO Annotator REST API. I spent some time looking at the documentation
provided at and
also found the Reccomender example R code. I managed to write some R code
to query the Annotator, anyway I've not been able to find a more clear
explanation of the tunable parameters. For example, when you specify
exact_match=FALSE how many mismatches in the word are allowed? Is there a
way to obtain matches for plurals or do I have to apply some stemming
before? why mappings are always empty even if i set the parameter ="all"?
Do you provide a way to exclude parent concepts where there is a more
specific match? (As an example, in a sentence containing the words "BRAIN
and "BRAIN" but I would like to exclude "EMBRYO" and "FIBROBLAST").
Below you can look at how I usually query the annotator. I would also like
to ask if the apikey that I have obtained can be used by an application,
since I'm writing an R package for bioconductor. 

  resultDF <- NULL		
  results_list <- NULL
  # Process each sentence
  for (sentence_id in 1:length(samples_sentences)) {
  		curr_sentence <-
as.character(samples_sentences[sentence_id]) # current sentence to process
		cat("#", sentence_id, "from", length(samples_sentences),
":", curr_sentence, "\n")  		
		JSONannotation <- postForm(REST_URL,
    		#"include" = "all"
    		"mappings"= "all",

I'm sorry for asking you all these questions by e-mail, but I didn't find
any alternative (even looking at the questions in the google group). Can
you please help me? Thanks,


Eugenia Galeota, PhD 
Center for Genomic Science of IIT at SEMM
Computational Epigenomics
Email: eugenia.galeota at
Tel: +39 02 9437 5046
Via Adamello 16, 20139 Milan, Italy

bioontology-support mailing list
bioontology-support at

More information about the bioontology-support mailing list