The hardest question in biology

Evolution explains how life diversified once it existed. It does not explain how it started, because evolution needs inheritance and inheritance needs something already copying itself. Getting from chemistry to the first thing that made copies of itself is the gap, and it is genuinely open.

The circle at the centre of it

In life as it exists now, the division of labour is strict. Nucleic acids carry information and are copied. Proteins do the work, including the work of copying nucleic acids. Neither can be produced without the other: the instructions for building proteins are held in nucleic acids, and reading those instructions requires proteins.

This is not a small chicken-and-egg puzzle. It is the reason the origin of life is hard. Any proposal has to explain how a system with two mutually dependent halves got started when neither half is useful alone.

The mutual dependency of information and machinery, and the RNA world proposal On the left, two boxes labelled information and machinery point at each other in a closed loop, each requiring the other, with no entry point. On the right, a single molecule labelled RNA does both jobs at once, storing information and acting as machinery, providing a starting point from which the two roles could later separate. The problem One proposed way in information machinery builds no way in RNA stores information and acts as machinery one molecule doing both jobs breaks the circle
The RNA world hypothesis does not claim RNA was first. It claims a stage existed where one kind of molecule did both jobs, after which the roles could specialise.

What the RNA world proposal gets right

The proposal rests on a genuine discovery rather than a convenience. RNA can store information like DNA, and some RNA molecules also catalyse chemical reactions, which was thought to be the exclusive business of proteins. Molecules that do both exist and have been studied for decades.

There is also a striking piece of circumstantial evidence sitting inside every living cell. The ribosome, the machine that builds all proteins, performs its central chemical step using RNA rather than protein. The most fundamental machinery in biology has an RNA core, which is what you would expect if RNA came first and proteins were added later.

What the proposal does not do is explain where RNA came from. Assembling its components under plausible early-Earth conditions is difficult, getting them to link up in the right way is harder, and doing it without a chemist arranging the conditions is harder still. Progress has been real and the gap is not closed.

The other serious lines

Two alternatives take the problem from different directions, and it is quite possible that the eventual answer draws on more than one.

Metabolism first. Perhaps self-sustaining cycles of chemical reactions came before any molecule that stored information, with replication arriving later on top of an already-running chemistry. Deep-sea vents provide the sort of persistent energy gradient and mineral surfaces this would need.

Compartments first. Simple fatty molecules spontaneously form enclosed bubbles that grow and divide without any machinery at all. A boundary keeps useful products from dispersing and makes competition between compartments possible, which is a precondition for anything like selection.

Nobody can agree where the line is

Underneath all of this sits an unresolved definitional problem: there is no accepted definition of life. Proposed criteria include metabolism, reproduction, response to environment, and the capacity to evolve, and every one of them either excludes things generally considered alive or includes things generally considered not.

Viruses fail most definitions and are clearly biological. Fire consumes energy and reproduces after a fashion. Crystals grow and template copies of themselves. The most useful working definition is probably a system that maintains itself and can evolve, and even that is a choice rather than a discovery. Any account of the origin of life is therefore partly an account of where we decided to draw a line.

Starting after the hard part

It should be stated flatly: this project does not address the origin of life, and it starts on the far side of the gap.

A universe here begins with organisms already present, already reproducing, already carrying an inheritable genome and a functioning controller. Everything difficult about the origin of self-replication has been handed to the world by construction. The founding population is placed there; it did not arise.

That is a normal position for artificial life work, and it is worth being clear that it is a limitation rather than a modest scope. The field's origins question, how a system that copies itself arises from one that does not, goes back to von Neumann's self-reproducing automaton and the insight that such a system needs a description of itself that it both copies and interprets, worked out before anyone knew that was how DNA is used. That is set out on cellular automata and artificial life.

The plain version

Nothing on this site is evidence about how life began, and no result here bears on the question in either direction. What runs here is what happens after inheritance already exists, which is a much easier problem and still produces plenty that is not understood.

Keep going

Watch the universe