DNA is a tangled molecule. In a cell, the double helix loops, coils, and knots itself into configurations that must be untangled before replication can proceed. Topoisomerase enzymes cut the DNA strand, pass another strand through the gap, and reseal the break — changing the knot type in a mathematically precise way.
The topological approach to enzymology treats the enzyme's action as a mathematical operation on knots. Incubate circular DNA with an enzyme, observe which knot types appear in the products, and work backward to determine the enzyme's mechanism. The enzyme leaves a topological signature — a specific pattern of knot type changes — that uniquely identifies its binding geometry and strand-passage mechanism.
Machine learning now recognizes and generates knotted molecular structures (arXiv:2501.12780), extending this analysis beyond DNA to protein knots and synthetic polymer knots. The knotted structures have practical value: their “special catalysis and extraordinary structural stability” arise precisely from the topological constraint that prevents the chain from untangling without being cut.
The through-claim: the knot isn't a defect — it's the functional element. A protein with a knotted backbone has enhanced stability because thermal fluctuations can't unknot it. A DNA molecule with enzyme-induced supercoiling has controlled access to its genetic information because the topology gates transcription. The topology IS the regulation.
This inverts the usual relationship between structure and function. In most molecular biology, function follows from three-dimensional shape — the lock-and-key model. In topological biology, function follows from connectivity — how the chain threads through itself. Two molecules with identical atomic coordinates but different threading patterns have different properties. The information is in the path, not the positions.