I find this stuff fascinating, but there is clearly no agreement yet. There is also this recent paper from Carter and WIlls, where they claim that both nucleic acid and proteins co-evolved together in the first life forms. They claim to have some experimental evidence that supports their hypothesis that I do not know enough about to evaluate the validity of their claim. So we have:
(1) RNA world - Nucleic acids came first - protein was added later. Many papers along these lines.
(2) The paper Greg cited - Protein came first, nucleic acid was added later.
(3) The paper I cited - The two co-evolved together. I like this paper in that they claim there were initially just two amino acids in the protein "alpahbet" (today there are 20), and more were added over time as complexity evolved.
I have some issues with this protein first approach compared with a previous thread (on the evolution of the ribosome and associated functions):
1) The model is not very detailed chemically (only 2 kinds of amino acids considered). Modern protein folding research/analysis/computer modelling is much more complex. Information is not being used (this chemical dumbing-down does not appeal to me).
2) Only an auto-catalytic sequences (of amino acids) would make more of themselves. Non-auto-catalytic sequences could catalyze the formation of other sequences, which could do other things, possibly even aid the catalytic sequence. Either way, producing an amino acid sequence based on the "reaction site" of a protein would seem to be limited to producing very specific sequences. it seems that the sequence at the reaction site would have to change to determine the reactions of a different sequence. Alternatively, the specificity of the sequences generated might be poor, which would allow a wider variety of sequences to be generated, but would also be less efficient. This would seem to be a big difficulty for this approach.
3) It is not clear to me that this underlies the origin of life (with other non-protein molecular stuff added later). Since the origin of life does not have an agreed upon definition, the line for a molecular system to cross to step into the world of the living is not defined. Instead, I see a more fruitful approach as looking for reasonable ways to assemble an interacting set of components that can:
self-replicate (their own information laden sequences)
create and maintain their own biochemical mini-environment (or local-environment), thus being able to control their replication and other processes
dependably harness some source of energy (probably environment at first) to get things done
All of these steps are interesting.
After achieving them all, a majority might agree the system would be alive.
4) The ribosome evolution paper (comparatively, a very detailed analysis) talks about the both primitive proteins and RNAs being present as does the Carter and Wills article mentioned by @phyzguy. This seems more reasonable to me since there should be uncontrolled and undirected production of biochemicals prior to the precursors being swept up after biological processes evolve.
As the ribosome complex was assembled, proteins glommed onto its surface, adding stability and possible chemical functions.
5) The ribosome paper also mentioned that the proteins that are generated (by the proto-ribosome) would pass through a tunnel penetrating the ribosome which kepts the amino acid chains linear rather than allowing them to fold over and react with themselves (cyclizing) which could results in a population of amino acid loops 2 (or a few) amino acids long. These could not be further extended to longer chains (the amino and carboxyl sites of all amino acids all being used). This would limit the average complexity to the amino acid chains generated. Assuming the chemistry of this is correct, this would seem to argue against a proteins only approach.
If something had to be first, I would always choose RNA.
Particularly considering the experiments of in vitro evolution of RNA in Szostak laboratories.
But, nonetheless, I think that the most plausible explanation is that two self-replicating molecules, RNAs and some simple crystal-like amino-acid sequences, combined together.
And personally, I am not sure that the first self-replicatin RNA-protein complex had to resemble ribosomes.
When comparing similarities of ribosomes between domains, that probably is the most likely
i would somehow give my advantage to signalling proteins, such are adenylyl cyclases or G-proteins.
I tend to be skeptical of unvalidated computations.
Protein molecules can build physical structures. so they must have been first, RNA cannot do that.
RNA can change proteins though.
Chicken or egg?
RNA can built "physical structures" (meaning I guess not primarily for information conveyance):
ribosomes (mostly RNA)
ribozymes (act like enzymes)
As carriers of information, of course, they also have some physical structure to themselves.
Separate names with a comma.