The evolution of life on Earth is a history of countless successes and failures. The variations arise through chance, but within physical, chemical, and biological organization and limitations. Natural selection acts on these variations, retaining forms that remain functionally viable, while many other variations persist through neutral evolution, shaped not by adaptive advantages but by drift and historical contingency. Quite frequently, evolution creates different solutions to the same problem, where different variations create their own rules. These rules are sometimes locally reliable, within model organisms for example, but not guaranteed to confine the full space of biological solutions. Comparative genomics pushes us from asking “is this true?” to “what other solutions exist?”. It helps us discover outliers; these outliers could act like anomalies in the paradigm, in Kuhnian terms. They don’t automatically overthrow the rule, but they expose where the rule is incomplete or where its scope is narrower than assumed. Our recently published work in Nature and Nucleic Acids Research represents an intriguing example of how non-model organisms, protists in our case, and comparative genomics can serve as a source of problematization that guides subsequent mechanistic elucidation of such variations.
Evolution: a generative grammar of viable mechanisms
Life is a highly structured and organized complex system bound to the foundational laws of physics and chemistry. Yet, life operates far from equilibrium, and this nonequilibrium organization makes biological processes intrinsically dynamic and historically contingent. Thus, life is not a static realization of physical law, but a fluctuation-sensitive system in which chance can be amplified into biological variations within physical and evolutionary constraints. The different context-dependent environmental and evolutionary forces act on these variations, filtering or promoting. Examples of such variations include mutations, gene/genome duplications, and recombinations along with others that are either eliminated by natural selection or maintained in certain contexts.
With that being said, evolution for life is somewhat similar to what generative grammar is for languages. In his seminal book Syntactic Structures, Noam Chomsky proposed the theory of “generative grammar” where he argued that languages are a finite set of formal, cognitive rules capable of generating infinite number of grammatically correct sentences. On the other hand, evolution by the processes of replication, inheritance, variation, and selection creates an open-ended space of diversity of life forms and of underlying diverse processes.
The problem of problematization
Molecular biology is an exceptionally powerful tool at establishing causal mechanisms in an incremental manner. It does so by working on a small number of tractable systems under highly controlled laboratory conditions. The resulting mechanistic statements are often correct and robust sensu stricto, that is, within a bounded experimental context, yet they can silently acquire a broader authority than they have the right to.
For decades, molecular biology in model organisms performed the puzzle-solving experiments within a highly advanced reductionist paradigm. According to the philosopher Karl Popper, scientific claims are considered stronger not by being treated as universally true, but by remaining open to refutation and by surviving severe tests. Evolution, by creating this open-ended space of molecular diversity, provides the comparative ground for biologist to test the limits of their mechanistic rules. The raison d’être of comparative genomics, in this context, is not simply to expand the catalog of biological variation, but to act as a source of problematization, reopening and questioning settled rules, exposing hidden assumptions, and revealing where apparently general mechanisms are in fact local solutions.
Comparative genomics: an engine of problematization
Biology is largely explained through mechanisms rather than axioms. Progress in biology often depends on identifying concrete explanatory problems. Problematization, here, is better illustrated in the moment when a supposedly general role is subject to extensive questioning of its scope and limits, thereby generating a new mechanistic question.
Comparative genomics analyzes the complete genomes sequences of different species to identify evolutionary relationships, conserved sequences, and genetic variations. Therefore, evolution creates a space of possibilities; comparative genomics samples realized variants across lineages. It reveals other solutions that nature created for a certain problem and sets the boundary conditions of extreme possible configurations.
By mining through sequenced genomes, comparative genomics help identify a signature; a gene, a structure, or a certain context. Then, the distribution is mapped where correlations and constraints are inferred. The detected signature helps in generating a minimal working hypothesis and in deriving testable predictions. Mechanistic tests can then be done in tractable systems. If the mechanistic experiments complement the bioinformatic analysis, the rules are updated with an explicit scope.
Why non-models matter: sampling the possibility space
In a comparative genomics move, colleagues and collaborators compared the tRNAs of the non-model parasite Blastocrithidia nonstop to those of closely related parasites. B.nonstop is a unicellular parasite that has reassigned all 3 stop codons into sense codons. They found that the parasite bears 2 fully cognate tRNAs-Glu that decode UAG and UAA, whereas the molecule implicated in UGA decoding is a rather peculiar tRNA-Trp with a CCA anticodon. Strikingly, B.nonstop tRNA-Trp differs from those of the other closely related parasites by having a shortened anticodon stem, a 4-base-pair AS compared to the canoncical 5-bp.
This raised a simple but destabilizing question: if the anticodon is unchanged, can a structural variation far from the anticodon, such as the 4-bp AS, alter decoding capacity of the tRNA and enable UGA decoding? Experiments in Kachale et al. 2023 proved that the shortening of the AS, by unpinning of the top base pair, can effectively enable C:A pairing at the wobble position and subsequent UGA reassignment in B.nonstop and the ciliate Condylostoma magnum.
We pushed the question even further, is this architectural deviation a lineage-specific curiosity or it can be generalized across various life forms? If it recurs, can it be framed as an extended superwobble hypothesis where the 4-bp AS=C:A wobbling? In Fakih et al. 2026, we again referred to comparative genomics at scale and analyzed a dataset of around 42k bacterial genomes. We found that the 4-bp AS tRNAs are widespread across bacterial phylogeny and occur across different tRNAs species. We also validated experimentally that these tRNAs are functional using Escherichia coli, showing that a 4-bp AS tRNA-Trp can promote UGA readthrough in a model bacterial system.
In this real-life practical example from our lab, we managed to use non-model protists and comparative genomics as a fuel for our experiments on the bench.
To summarize, evolution creates a constrained yet open-ended possibility space of viable molecular mechanisms, and comparative genomics is the tool we use to empirically sample the space, making rare solutions visible. These boundary cases become problematization triggers that turns rules into scoped hypotheses with testable predictions. Our discovery workflow in the B.nonstop tRNA story illustrated how a comparative signal can be tested mechanistically and lead to an updated scope and framework.
Best regards,
Fadel
References:
Fakih, F. et al. Frequent occurrence and predicted functions of tRNAs with 4-base-pair anticodon stems in bacteria: extended superwobble hypothesis. Nucleic Acids Res. 54, gkag327 (2026). https://doi.org/10.1093/nar/gkag327
Kachale, A. et al. Short tRNA anticodon stem and mutant eRF1 allow stop codon reassignment. Nature 613, 751–758 (2023). https://doi.org/10.1038/s41586-022-05584-2
Kuhn, T. S. The Structure of Scientific Revolutions (University of Chicago Press, Chicago, 1962)
The featured image was generated by ChatGPT.
RNA Stuff 