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    

[Gofriends] Evidence Code Ontology (ECO) new release

Marcus Chibucos mchibucos at som.umaryland.edu
Thu Jan 20 14:45:05 PST 2011


To Evidence Code Ontology users:

The Evidence Code Ontology (ECO; evidence_code.obo) has undergone 
substantial revision over the past year, and this release contains many 
significant changes, outlined below. The present release is the first of 
several to come over the next several months in our ongoing effort to 
improve ECO with respect to ontology structure and logic, definitions, 
and user requests. Here we have attempted to tackle the major and most 
pressing issues, although further changes, such as modifying definitions 
of more granular terms and addressing user requests will continue.

1. Removal of "inferred from" and cleaning definitions:
In its prior form, ECO was inconsistent with respect to whether it was 
an ontology of evidence types, inferences, or experiment. We believe it 
should be an ontology of evidence types, and have modified terms and 
structure accordingly. Consequently we removed the "inferred from" text 
from the term names (all of the evidence types are used in the act of 
inference, and including that text in each term name is essentially 
meaningless). We have maintained the former "inferred from X" names and 
three letter codes as synonyms. We have also cleared references to GO 
annotation practices in definitions (this is still a work in progress). 
Many of the details previously found in the definitions can now be found 
in the comments.

2. Introduction of "assertion method" and cross product terms:
In this significant change, the ontology now has two roots, an evidence 
type and an assertion method. The evidence type forms the backbone of 
the ontology, where each evidence term describes a type of information 
used to support an assertion (evidence is now defined). The assertion 
method is the means, either automatic or manual, by which a statement is 
made about an entity. For all terms in the ontology, cross products of 
manual assertion and evidence type will be made (at present this is not 
complete). Where appropriate, automatic cross products will be 
completed, as well. Annotation should be performed using the cross 
product terms.

3. Renaming of IEA, RCA, and other terms:
The GO evidence code "inferred from electronic annotation" has been 
renamed "automatic assertion" (and synonyms for the prior term made). A 
new group of terms was created to reflect "combinatorial analysis". In 
silico analysis was merged into "combinatorial analysis" and "reviewed 
computational analysis" (RCA) was renamed "computational combinatorial 
analysis" and made a child term.

4. Obsoletion of "not_recorded":
We made NR obsolete, as it is not a type of evidence.

5. New evidence terms:
We have added the three new "NOT" evidence codes. We have constructed 
their names and definitions to conform to the guidelines we have 
established for ECO. They are children of the term "phylogenetic 
evidence", itself a child of "similarity". We have added three-letter 
codes as synonyms. We have added a number of other evidence types in 
response to user requests, such as terms that describe: various 
sequencing types, chromatin immunoprecipitation, PCR, chromatin 
conformation capture, anatomical perturbation, morpholino experiment, 
immunolocalization, and so forth.

6. Future work:
We will continue to clean up terms and definitions in future upcoming 
releases, focusing in particular on more granular term definitions. We 
will also add the full complement of the manual assertion method by 
evidence type cross products; we will add automatic assertion method by 
evidence types in response to user requests. We will continue to map ECO 
terms to the Ontology for Biomedical Investigations (OBI).

Please post comments, term requests, and questions on the SourceForge 
tracker at http://sourceforge.net/tracker/?group_id=36855&atid=1127740 . 
We are happy to provide additional details on the changes; please 
contact us if you have any questions.

Best regards,
Marcus Chibucos and Michelle Giglio
Institute for Genome Sciences



More information about the go-friends mailing list