Reactions and Transformations - Design and Use of Relational Databases in Chemistry - page 106

NH 2

O

O N

O

HN

HO

N

O

Figure 9.2 Four-component Ugi reaction.

In this case, reactant atoms that do not actually appear in the product are
simply not mentioned in the product side of the SMIRKS. This is a useful
feature of SMIRKS, allowing some reactant atoms to “disappear.” This is
convenient with hydrogen atoms, water oxygen atoms, and sometimes
with other atoms. Depending on how the product SMILES will be further
processed, it may be desirable to use this feature in order to retain only
the main product. In addition, an unmapped hydrogen atom appears
in the product on the [N:40] atom coming from the isocyanide. It is not
important, and indeed probably not clear from the mechanism, exactly
which hydrogen atom this is. For the purposes here, it is not crucial to
exactly map that product hydrogen atom to a particular reactant hydro-
gen atom.
If the before-mentioned SMIRKS is stored in the smirks_lib table, then
the following SQL

select xform(
ARRAY['CCC=O','CN','c1ccccc1C(=O)O','[C-]#[N+]C1=CCCCC1'], smirks)
from smirks_lib where name='Ugi4CC';

will return CCC(N(C)C(=O)c1ccccc1)C(=O)NC2=CCCCC2. This transfor-
mation can be drawn as shown in Figure 9.2.

9.4 Canonical Reaction SMILES
Reaction SMILES is a syntactical extension of regular SMILES. Since it
is composed of regular SMILES punctuated with periods and the >>