Chapter 27

Drug Discovery

Presently some 4000 medicinal drugs are approved for use in the world; there are

about 22,000 human genes, and many more gene products. This suggests that there is

as yet considerable scope for the discovery of more drugs, neglecting the possibility

that some drugs may act on multiple gene products. Indeed, avoiding such undesirable

side reactions is a major difficulty in discovering clinically safe drugs.

Formerly drug discovery was largely a matter of serendipity, as epitomized by

the well-known story of Alexander Fleming and penicillin. The huge technological

advances of molecular biology, biochemistry and bioinformatics have made it pos-

sible to rationalize drug discovery to such an extent that the process can be largely

automated. Just as an interactive genetic algorithm can be more efficient than a

genetic algorithm, drug discovery also benefits from intelligent human intervention,

however.

The first task is to discover drug targets. These are either genes, or gene promoter

sites; or proteins. ¹ A drug can activate or deactivate. A molecule that binds almost

irreversibly to a gene or its promoter site would be predicted to inactivated it; similarly

with one that binds to the active site of an enzyme, or a site that interfaces with a

protein’s binding partner.

We have already noticed that it is much easier to sequence a gene than sequence

a protein; indeed now that the entire human genome has been at least sequenced a

canonical version of all human genes is in principle available—one does not always

know whether a particular sequence of bases is a gene or something else. But if

it can be identified as a gene, this implies a polypeptide sequence, from which it

may be possible to calculate its three-dimensional (tertiary) structure. That in turn

will provide clues to its possible interactions with other genes or gene products. The

corpus of knowledge known as QSAR (quantitative structure—activity relationships)

may allow further deductions regarding enzymatic activity and the like.

1 Non-coding regulatory RNA could also be a target.

© Springer Nature Switzerland AG 2023

J. Ramsden, Bioinformatics, Computational Biology,

https://doi.org/10.1007/978-3-030-45607-8_27

365