periodic table

[muthuvenkat]

perhaps I am being dense, but I can't reconcile my (weak) knowledge of basic chemical elements and the representation of the periodic table in chebi

All elements in chebi follow roughly the same organisation:

is_a CHEBI:33679 helium molecular entities +is_part_of CHEBI:30217 helium ++is_a CHEBI:30218 helium-3 [DEF: "The stable isotope of helium with relative atomic mass 3.016029. The least abundant (0.000137 atom percent) isotope of naturally occurring helium."] ++is_a CHEBI:30219 helium-4 [DEF: "The stable isotope of helium with relative atomic mass 4.002603. The most abundant (99.99 atom percent) isotope of naturally occurring helium."] ++is_a CHEBI:37003 helium-6 [DEF: "The radioactive isotope of helium with relative atomic mass 6.01889 and half-life of 806.7 ms."] ++is_a CHEBI:37004 helium-8 [DEF: "The radioactive isotope of helium with relative atomic mass 8.03392 and half-life of 119.0 ms."] +is_a CHEBI:33680 elemental helium +is_a CHEBI:33691 helium hydrides

What is the difference between helium and elemental helium?

What is helium molecular entities, and in what sense is helium part_of helium molecular entities?

[sorry, i know I've said it multiple times before, but the use of plurals really grates here...]

It would seem to be relatively easy to automatically populate this part of chebi with rigorous necessary-and-sufficient definitions. This can be done as both text definitions and obo definitions (obo-format has no problems saying things like "carbon has_part (6) proton" and "carbon-14 has_part(14) nucleon" - this will also get represented in the corresponding owl; many people will find this VERY useful)

Reported by: cmungall

[muthuvenkat]

Logged In: YES user_id=802919 Originator: NO

> What is the difference between helium and elemental helium?

'helium' is an element. Element is a type of atom which is defined by number of protons in nucleus. Element is not the same as a molecular entity.

'elemental helium' is any molecular entity that consists of helium atom(s) only. As you can see, there are monoatomic (e.g. helium(0)] and diatomic [e.g. dihelium(2+), dihelide(1-)] species of helium.

> What is helium molecular entities

molecular entities that contain helium (*).

> and in what sense is helium part_of helium molecular entities?

if you reverse the definition (*), you will get the answer yourself, i.e. that Helium (element) is part of any helium molecular entity

For example: hydridohelium(1+) (CHEBI:33688) is_a helium hydrides (CHEBI:33691) is_a helium molecular entities (CHEBI:33679)

thus, helium (element) is part of hydridohelium(1+)

> (obo-format has no problems saying things like "carbon has_part (6) proton" and "carbon-14 > has_part(14) nucleon" - this will also get represented in the corresponding owl;

So far, we do not have a mens to represent this within the ChEBI structure. We'll need to change structure of some tables within the db to accomodate attributes (like numbers). OBO is just one of our output formats. Alternatively, one can think of representing this particular kind of information with nuclear formulae like this:

np (deuterium) n2p (tritium) n6p6 (carbon-12) n8p6 (carbon-14) etc

Kirill

Original comment by: kiri11

[muthuvenkat]

Logged In: YES user_id=1473024 Originator: NO

It's very common in atomic spectrometry papers to test for the presence of a particular element in a sample. And frequently people talk about elements without specifying their exact chemical context, like this:

"Because Se and sulfur (S) are same-group elements, plants having high S content are expected to have high Se assimilating ability."

So, from your answer, should we annotate those mentions of elements as "selenium" rather than "elemental selenium"? And what about phrases like "nitrogen cycle"?

Original comment by: batchelorc

[muthuvenkat]

Logged In: YES user_id=802919 Originator: NO

> hould we annotate those mentions of elements as "selenium" rather than "elemental selenium"? > And what about phrases like "nitrogen cycle"?

These both refer to the elements, not molecular entities. So yes to "selenium" and "sulfur" (in atomic spectrometry sense) yes to "copper-63" (in EPR sense) yes to "nitrogen" in "nitrogen cycle" ("elemental nitrogen", more precisely dinitrogen, is only one participant of nitrogen cycle)

Kirill

Original comment by: kiri11

[muthuvenkat]

Logged In: YES user_id=254723 Originator: YES

>> What is the difference between helium and elemental helium?

> 'helium' is an element. Element is a type of atom which is defined by > number of protons in nucleus. Element is not the same as a molecular > entity.

I'm a bit confused already. You have introduced a term Element, saying "Element is a type of atom". But in ChEBI, element is a related synonym for atom(s). Can you substitute "atom" for "element" in the sentence above? If so "element" should be an exact synonym for "atom". If not, then why use the term "element" in your explanation if it is relegated to a related synonym?

Also, when you say "element is a type of atom which..." it makes it sound as if "element" should stand in an is_a relation to "atom".

I don't know how to interpret "Element is a type of atom which is defined by number of protons in nucleus". Can you give me an example of an atom that is not an element?

You say element is not the same as molecular entity. Yhe definition should tell us why this is the case:

molecular entities [DEF: "A molecular entity is any constitutionally or isotopically distinct atom, molecule, ion, ion pair, radical, radical ion, complex, conformer etc., identifiable as a separately distinguishable entity."]

I don't think ontology definitions should contain words like "identifiable" or "distinguishable" as this lends them a subjective air. Distinct from what?

I know this definition comes from a respectable terminology of chemistry, but I don't think ChEBI should shy away from providing ontologically sound definitions if it is indeed an ontology and not a terminology.

From the above is sounds as if certain atoms can be molecular entities, in and of themselves? Is this intentional?

> 'elemental helium' is any molecular entity that consists of helium atom(s) > only.

Is the (s) to indicate that [1] 'elemental helium' can consist of a single helium atom or [2] ChEBI calls an "atom" an "atoms"?

If the former, then a single atom of 'helium' also an instance of 'elemental helium'? If not, what is it in your definition of 'elemental helium' or the parent term 'molecular entities' that rules this out?

> As you can see, there are monoatomic (e.g. helium(0)] and diatomic > [e.g. dihelium(2+), dihelide(1-)] species of helium.

You're introducing a new relation "species_of" which is not defined in ChEBI. Of course, every chemist knows what this means, but part of the point of an ontology is to provide definitions of terms using simpler terms that are themselves defined in the ontology (or some related ontology).

Also, you say "'helium(0)' is a monoatomic species of 'helium'", but I don't see how this relationship is represented in ChEBI. ChEBI says:

is_a CHEBI:33679 helium molecular entities +is_a CHEBI:33680 elemental helium ++is_a CHEBI:33315 monoatomic helium +++is_a CHEBI:33681 helium(0)

Helium is in a separate branch under "atoms".

I'm also not sure how 'helium(0)' differs from 'helium'

>> What is helium molecular entities > molecular entities that contain helium (). >> and in what sense is helium part_of helium molecular entities? > if you reverse the definition (), you will get the answer yourself, i.e. > that > Helium (element) is part of any helium molecular entity

If you are using part_of in the OBO/RO defined way, then you are saying "all instances of helium (element) are part_of some helium molecular entity". All relations in obo should be in the all-some direction. So if you want to say all helium molecular entities have helium as part then you would say:

helium molecular entity has_part elemental helium

> For example: > hydridohelium(1+) (CHEBI:33688) is_a helium hydrides (CHEBI:33691) is_a > helium molecular entities (CHEBI:33679) > thus, helium (element) is part of hydridohelium(1+)

I'm afraid I don't follow this at all.

Regardless of my understanding or lackof, you are interpreting the obo format in your own way contrary to the spec which has implications for interoperability.

What are the exact deductions you used to arrive at the conclusion above?

I think if you add a has_part link then you can infer that hydridohelium(1+) has_part helium (element)

>> (obo-format has no problems saying things like "carbon has_part (6) >> proton" and "carbon-14 >> has_part(14) nucleon" - this will also get represented in the >> corresponding owl; > So far, we do not have a mens to represent this within the ChEBI > structure. We'll need to change structure of some tables within the db to > accomodate attributes (like numbers). OBO is just one of our output > formats.

By exporting obo you get owl for free, and these are the two main formats for ontologies.

> Alternatively, one can think of representing this particular kind of > information with nuclear formulae like this: > > np (deuterium) > n2p (tritium) > n6p6 (carbon-12) > n8p6 (carbon-14)

If there is a formal grammar for this then we can convert to obo/owl

Sorry if I'm missing some subtleties here. I have a very simple view of ontologies and of chemistry. In my simple world view, every instance of a chemical entity in reality could be placed into a single class based on definitions of the form "an X is_a Y that Z", where Y is a term in the ontology and Z is one of:

- a constraint on what parts X has

• a constraint on what X is part of
• a constraint on what X is otherwise physically related to (via bonds, etc)
• a quality that is possessed by X (such as charge, or possession of mass)

With names that reflect the definitions

but obviously there's a lot I'm missing!

> Kirill

Original comment by: cmungall

[muthuvenkat]

Logged In: YES user_id=802919 Originator: NO

Chris, your post is just toooo long to be responded in this place point by point. Really.

In ChEBI there is no difference between "element" and "atom". This difference however existed until XIX century because not everybody has accepted atomic theory. So, in ChEBI context hydrogen (element) is exactly the same as atom of hydrogen. And atom of hydrogen is an abstract concept, which may refer to anything, from free proto/deuteron/triton to the hydrido group within a molecule. Similarly, helium (element) is exactly the same as atom of helium. We make the distinction between abstact concept (element helium) and any molecular entity that includes helium, even if the most familiar entity is a monoatomic helium. It may be free The only complication is that free helium

Now we gave CHEBI:33250 name "atoms" rather than "elements" because many people use "elements" in a different (older) meaning, which is that of elemental substance, or confuse both of these meaning. And we wanted to avoid this confusion. Atoms (=elements) are parts of molecular entities, not molecular entities theselves.

Original comment by: kiri11

[muthuvenkat]

Logged In: YES user_id=254723 Originator: YES

You say that "atom of hydrogen" can refer to a "deuteron". Do you mean "refer to" in informal colloquial usage, or in ChEBI? Presumably the latter, as the relation between the two terms in ChEBI is "is_part_of".

I don't think there's any need to bring abstract concepts into it - that just confuses things. ChEBI should say say what the term means, independently of any abstract concepts. Especially if the distinction between helium (element) and "helium molecular entities" (which subsumes "elemental helium"!) is due to the fact one is an abstract concept.

Would you object to the following rewriting of your sentence: "We make a distinction between an atom of helium and any molecular entity that includes helium"? This makes more sense to me. Though I am not sure if you use the word "includes" in the sense of "part_of", or if you mean included within the category or concept of helium. (This is why it's important to eliminate concept talk and only talk of real things - otherwise these use/mention confusions will be rampant).

I will take the former reading: there is a distinction between an atom of helium (a kind of atom) and a diatomic helium molecule (a kind of molecule, which has atoms as parts). This makes sense to me.

But I'm not sure how monoatomic helium fits in. In ChEBI this is not a kind of (is_a) helium atom. I can't get my head around this.

Of course, there is a difference between an atom in it's isolated monoatomic state and its di/tri/-atomic state. This is what the is_a relation is for:

monoatomic helium atom is_a helium atom helium atom in diatomic molecule is_a helium atom diatomic helium molecule has_part{2} helium atom

I can't see how isolating a helium atom makes it any less a helium atom, if it exhibits the properties that define a helium atom

Sorry if the sf tracker isn't the best place for a discussion - but chebi doesn't have a public mail list like the other OBO ontologies.

Thanks for the clarification re elements. I only asked because you used the term "element" in your previous explanation.

Original comment by: cmungall

[muthuvenkat]

In the light of discussions today between the ChEBI team and Chris and an agreement that the best solution is to separate out elements from the main ontology and treat these as attributes of the atoms, this request is being moved into Pending while plans for progressing this are formulated.

Original comment by: mennis

[muthuvenkat]

• status: open --> pending

Original comment by: mennis

[muthuvenkat]

This Tracker item was closed automatically by the system. It was previously set to a Pending status, and the original submitter did not respond within 14 days (the time period specified by the administrator of this Tracker).

Original comment by: sf-robot

[muthuvenkat]

• status: pending --> closed

Original comment by: sf-robot

[muthuvenkat]

• status: closed --> pending

Original comment by: pauladematos

[muthuvenkat]

This Tracker item was closed automatically by the system. It was previously set to a Pending status, and the original submitter did not respond within 365 days (the time period specified by the administrator of this Tracker).

Original comment by: sf-robot

[muthuvenkat]

• status: pending --> closed

Original comment by: sf-robot