Reactions and Transformations - Design and Use of Relational Databases in Chemistry - page 100

CC(=O)O.CN>>CC(=O)NC.O and matches('CC(=O)O.CN>>CC(=O)
NC.O','CC(=O)O') returns a true value. The SQL
Select rxnsmiles From a table Where matches(rxnsmiles,'CC(=O)O');
selects all the reaction smiles that involve acetic acid (or its derivatives)
from a table that stores reactions.
Reaction SMILES represents a specific reaction between specific reac-
tants yielding specific products. As such, it can be very useful to store a
library of reactions of interest. These might be a record of reactions that
have been carried out at a company, a set of reaction plans in an academic
research group, or even a set of hypothetical reactions that might never
succeed in the laboratory. There is another extension of SMILES called
SMIRKS that is more general and can represent a class of reactions.
SMIRKS mixes elements of SMARTS with SMILES. A more complete
definition of SMIRKS is available at daylight.com. 1 Several authors discuss
the graph theory underpinnings of SMIRKS. 2 The reaction SMILES above,
CC(=O)O.CN>>CC(=O)NC.O, could be generalized to allow acid chlorides
as well as carboxylic acids. This would be CC(=O)[O,Cl].CN>>CC(=O)
NC.[*]. The use of the SMARTS [O,Cl] allows oxygen or chlorine. The sec-
ond product, which was water in the initial example, is expressed simply
as [*] since it could be either Cl or O. When SMIRKS is used, it cannot be
considered a type of SMILES and therefore cannot be used in the matches
function or other functions requiring SMILES. However, SMIRKS can be
used to transform a given set of SMILES according to the rule specified in
the SMIRKS: in other words, to carry out the transformation in silico.
9.3 Transformations
The word reaction is typically used to represent a specific reaction as well
as a general transformation. In this chapter, the word reaction is used to
mean a specific reaction represented using reaction SMILES. The word
transformation is used to indicate a change in a set of reactants to products.
SMIRKS is the language used to specify precisely how this transforma-
tion is to be carried out.
In order to carry out a transformation, it is necessary to know pre-
cisely which bonds are to be broken and which are to be made. While
this information is implicit in the SMIRKS above and can be understood
by any chemist, a more specific set of instructions is necessary in order
to make the transformation possible using a computer. Numbering the
atoms of the reactants and the corresponding atoms in the products accom-
plishes this. This produces an atom mapping. Many sketching programs
can do this automatically or with some additional input from the user.
Using the example shown previously, an atom-mapped SMIRKS would