FAQ #18:
"ChEBI displays synonyms and formulae exactly as they
are portrayed in the source databases."
I think we have to revisit our policy re. synonyms that
are manually edited in ChEBI. My view is that we do not
have to reflect all the limitations of the sources if
we can add value by formatting, adding semantically
meaningful tags etc. Nor shall we change the source of
the terms modified in such way to "ChEBI" as in this
case we will lose track from where we did take the synonym.
Examples of formatting that I consider are:
1) converting "index names" into "normal names"
(especially, when by leaving them "as is" will lead to
wrong bracket nesting)
2) removing the part of "term" that actually is the
source of data
(example: "Acidum penteticum" [INN-Latin] => we take
conly "Acidum penteticum")
3) de-capitalisation (Pentetic acid -> pentetic acid)
4) adding various XML tags, even if they do not change
the appearance of the term (e.g. replacing actual
symbol for alpha in NIST Webbook term with
<stereo>alpha</stereo>)
All of the above and more is permittable as long as we
do not attribute our own terms to the other sources.
Kirill
Logged In: YES
user_id=1174860
I think a difficulty might arise if a user tries to search
for a ChEBI-curator-modified term within the source from
which it is supposedly taken. For example there is an entry
in ChemIDplus for 'Coenzyme A, S-butyl-' which has as its
only synonym 'Butyl coenzyme A'. If a ChEBI curator were to
change the former name to 'S-butyl-coenzyme A', anybody
using this altered name as a search term within ChemIDplus
would not find the entry. I think we need to find out
whether this sort of thing would pose a real problem to
users or not.
Marcus
Logged In: YES
user_id=802919
Originator: YES
After the discussion, the decison has been taken to implement a new type of synonym to reflect the fact that we are adapting them for ChEBI style. However it poses a representation problem since until now the type of synonym has been not shown in external ChEBI. I suggest the following phrasing
SYNONYM ADAPTED FROM
SYNONYM AS APPEARS IN
and for the representation, we will have to squeeze the synonym type between the term itself and the source, so it will look like:
Synonyms Synonym type Sources
(R)-Prunasin as appears in KEGG COMPOUND
D-prunasin adapted from ChemIDplus
and so on.
Comments?
Kirill