What is chemical space?

Chemical space is the conceptual collection of every molecule that could exist. Even when it is restricted to small, stable, drug-like organic molecules, a frequently cited estimate puts its size at around 10⁶⁰ compounds — vastly more than the number of molecules ever synthesised (on the order of 10⁸ registered substances).

Mapping a space too large to list

Researchers have enumerated parts of it. The GDB-17 database, for example, lists about 166 billion molecules containing up to 17 heavy atoms of C, N, O, S and halogens. Each extra atom multiplies the count, so complete enumeration of drug-sized molecules is impossible.

To reason about chemical space, chemists describe molecules with numbers — molecular weight, lipophilicity, polar surface area, or fingerprints — and treat each molecule as a point. Molecules close together tend to have similar properties, which is why similarity search works.

How compound libraries sample it

• Stock libraries — physically available compounds, typically hundreds of thousands to a few million, deliverable quickly.
• Make-on-demand (virtual) libraries — molecules designed from validated reactions and available building blocks. They reach billions of compounds, with synthesis taking a few weeks.
• Focused libraries — sets designed for a target class or a region of space, such as kinase inhibitors, covalent fragments or macrocycles.

Why diversity and novelty matter

Historic screening collections over-represent flat, aromatic, easy-to-make molecules, and many difficult targets — protein–protein interactions in particular — have no good binders there. Exploring under-sampled regions, such as larger three-dimensional ring systems, increases the chance of finding a hit where conventional libraries fail. That is one of the main motivations behind macrocycle libraries.