You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Recognize the equivalence of circularity and crosslinks which generate circular molecules. #109
Semantically, I think its useful to have a special syntactic sugar for circularity. Semantically, this connotes different meaning than a crosslink.
More generally, BpForms is not founded on the principle that there should only be 1 unique way to describe a polymer.
For circular polymers, there isn't necessarily a well-defined start position
There's nothing that prevents the community from defining monomeric forms which are themselves polymers. If such a monomeric form is semantically meaningful on its own and has a commonly accepted id or name, I don't see a problem with this.
The main goal of BpForms is to be concrete. I don't think this requires a unique representation of each molecule.
For circular polymers, there isn't necessarily a well-defined start position
I noticed that this a problem where the bpforms equality check fails. I had a brief look at the code, but the way it is structured I couldn't find a trivial fix.
changed the title [-]The `circular` attribute is redundant to crosslinking[/-][+]Recognize the equivalence of circularity and crosslinks which generate circular molecules.[/+]on Oct 5, 2020
Activity
jonrkarr commentedon Oct 5, 2020
Semantically, I think its useful to have a special syntactic sugar for circularity. Semantically, this connotes different meaning than a crosslink.
More generally, BpForms is not founded on the principle that there should only be 1 unique way to describe a polymer.
The main goal of BpForms is to be concrete. I don't think this requires a unique representation of each molecule.
paulflang commentedon Oct 5, 2020
OK, thanks. This all makes sense.
I noticed that this a problem where the bpforms equality check fails. I had a brief look at the code, but the way it is structured I couldn't find a trivial fix.
jonrkarr commentedon Oct 5, 2020
Yes, that a good point. We should change the code to recognize the equivalence of this special case. This comes up in a couple places
[-]The `circular` attribute is redundant to crosslinking[/-][+]Recognize the equivalence of circularity and crosslinks which generate circular molecules.[/+]