SMARTS pattern to match rings

[jeanmartel]

Hello,

I am trying to understand how to write SMARTS patterns to match rings in a structure. I am particularly confused about matching fused vs unfused rings. So for instance, a pattern that would match a pyridine ring in the structure but not a quinoline ring, vs a pattern that will return a hit if the structure had a pyridine or quinoline. Thanks in advance!

[s.roughley]

By default, the SMARTS atoms will match any number of rings, so a very simple SMARTS matching pyridine would be:

n1ccccc1

If you want to prevent any ring fusions, then you need to set the ‘R’ on each atom to 1, e.g.

[nR1]1[cR1][cR1][cR1][cR1][cR1]1

This is specifying that each atom is in exactly 1 ring - i.e. no fusion is possible.
If you want to allow fusion in a specific position, then you will either have to create separate SMARTS for each fusion position:

n1c[cR1][cR1][cR1][cR1]1  - Allows fusion on the N-C bond (i.e. 1,2- or 'a' or 'f')
[NR1]1cc[cR1][cR1][cR1]1  - Allows fusion on the 2,3- C-C bond ('b' or 'e')
[NR1]1[cR1]cc[cR1][cR1]1  - Allows fusion on the 3,4- C-C bond ('c' or 'd')

If you want to force fusion, then you need to specify ‘R2’ in those positions:

[nR2]1[cR2][cR1][cR1][cR1][cR1]1  - Forces fusion on the N-C bond (i.e. 1,2- or 'a' or 'f')
[nR1]1[cR2][cR2][cR1][cR1][cR1]1  - Forces fusion on the 2,3- C-C bond ('b' or 'e')
[nR1]1[cR1][cR2][cR2][cR1][cR1]1  - Forcesfusion on the 3,4- C-C bond ('c' or 'd')

If you want a single SMARTS that will give you a pyridine that must be fused somewhere, then you will need to use recursive SMARTS, e.g.

[$([nR2]1[cR2][cR1][cR1][cR1][cR1]1),$([nR1]1[cR2][cR2][cR1][cR1][cR1]1),$([nR1]1[cR1][cR2][cR2][cR1][cR1]1)]

NB the last SMARTS will only match a single atom, the ‘N’ atom of the pyridine ring - all the others will match the whole pyridine ring.

https://www.daylight.com/dayhtml/doc/theory/theory.smarts.html

Steve

[jeanmartel]

Thanks so much for the thorough explanation. It’s much clearer to me now.

Best,
Jean