Permanent generic parthood desired

The current ternary relations 

   r-at-all-times
   r-at-some-times

express permanent-specific relations. 

...

Being so, the pattern that has been termed as "Permanent relatedness" = at any 
time there is some instance of B related by r. (e.g. all cells contain some 
triglyceride) can only be expressed by the relation r-at-some-times (because it 
is not always the same triglyceride molecule). 

The relation r-at-all-times would then only be used for the stricter case 
("permanent specific relatedness") where there is not only always a relation to 
some instance of the same type, but always to the same entity (e.g. all cells 
contain some membrane - always the same)

The problem I see then is that r-at-some-time is never transitive, even if r is.
Example: "has-part-at-some-times"

How do I get the following inference in OWL:
- At any time all cells have some triglyceride molecule as part
- At any time all triglyceride molecules have some fatty acid residue as part
==> 
At any time all cells have some fatty acid residues as part

We cannot use the (transitive) relation "has-part-at-all-times" here, due to 
the fact that the molecules are not permanently part of a cell.  
Using the non-transitive relation "has-part-at-some-times" I don't get the 
inference I need.

Original issue reported on code.google.com by alanruttenberg@gmail.com on 15 Jun 2012 at 5:22

zhengj2007 / bfo-export

Permanent generic parthood desired #54