3 Rules and Inference
Rules and inference are important for tackling the challenges in bioinformat-
ics. For example, consider the Biomolecular Interaction Network (BIND).
The problem of defining interactions is very complex, and interactions must
be obtained from several sources, such as the Protein Data Bank (PDB), met-
abolic/regulative pathways, or networks. Rules can be used to model and
query these interaction networks.
3.1 Introduction to Rule-Based Systems
Anenzymeis a protein that exerts an effect on a specific molecule called its
substrate. The enzyme and its substrate combine to form theenzyme-substrate
complex. The active site of the enzyme and the corresponding part of the sub-
strate have shapes that match each other in a complementary manner, known
as the lock-and-key model, proposed by Emil Fischer in 1890. The substrate
fits into the enzyme just as a key fits into a lock. This is a good model for
visualizing how enzymes catalyze reactions. More recently, in 1958, Daniel
E. Koshland, Jr. described how the active site of an enzyme can change when
the substrate binds to the enzyme. Thus enzymes can take an active role in
creating a shape into which the substrate fits. This process is known as the
induced-fit model.
Aregulatory transcription factoris a protein that exerts an effect on the rate at
which a gene is transcribed. Such a protein forms a complex with a chromo-
some at specific binding sites determined by the sequence of bases at these
sites. One may consider the binding site of a transcription factor to be a
“word” encoded in the DNA. The “words” are the DNA binding motifs for
their respective transcription factors. The binding sites for a gene encode the
mechanism by which cells control such important biological functions as cell
