Uniqueness of tags?

Ralph Gauges ralph.gauges at eml.villa-bosch.de
Thu Aug 31 03:11:46 EST 2000


> 
> I would have
> 
> ?Reaction Substrate ?Compound XREF Reaction
> 
> ?Compound Type UNIQUE Peptide
>                       Nucleic_acid UNIQUE DNA
>                                           RNA
>                       Small_molecule
>           Reaction ?Reaction XREF Substrate
> 
> Then define subclasses of Compound as follows in subclasses.wrm
> 
> Class Peptide
> Is_a_subclass_of Compound
> Filter Peptide
> 
> Class Nucleic_acid
> Is_a_subclass_of Compound
> Filter Nucleic_acid
> 
> Class DNA_compound
> Is_a_subclass_of Nucleic_acid
> Filter DNA
> 
> etc.
> 
> Does this do what you want?  Note that I say DNA_compound because DNA
> is a reserved class.
> 
> Richard

If I would add all the extra attributes that e.g. peptide
has compared to Compound, it might work.
If I get this right, it might look something like
 ?Reaction Substrate ?Compound XREF Reaction
 
 ?Compound Type UNIQUE Peptide  Sequence TEXT
                                Function TEXT
                                ....
                       Nucleic_acid Type UNIQUE DNA
                                                RNA
                                    Sequence TEXT
                                    ...
                       Small_molecule ...
           Reaction ?Reaction XREF Substrate

To be honest, I have not looked at subclasses in detail yet.
But I assume that the filters could also be made more
restrictive. Like Filter Type.DNA instead of just DNA. But
again, what if I have subclasses of subclasses and so on.
This would get very complicated if I put everything in the
Compound model.
But this mailinglist is no substitute for reading the
documentation, so I don't want to bother you with more
questions that I might get answers to from the manual. I
think I will read the stuff about subclassing. Thanks again
for putting me on this track, maybe this will help me to
find the optimal solution for the problem.

Ralph








More information about the Acedb mailing list