A recent paper in the prestigious journal Science by Andrei Papkou and colleagues presents a fascinating, high-resolution map of an evolutionary “fitness landscape.” By creating and testing over 260,000 variants of a single bacterial enzyme, they sought to understand how populations navigate from lower-fitness states to higher-fitness peaks. The study, “A rugged yet easily navigable fitness landscape,” is being heralded as a powerful demonstration of Darwinian evolution’s ability to overcome obstacles. The authors found that even though the landscape of possible mutations is “rugged” (full of dead ends and sub-optimal peaks), the highest and most desirable fitness peaks are surprisingly easy to reach.
However, a closer look at the study’s methods and actual findings reveals a profoundly different story. Far from demonstrating the creative power of unguided evolution, this experiment masterfully illustrates the principle of adaptive degeneration. It shows how a pre-existing, information-rich biological system can be broken in specific ways to achieve a short-term survival advantage, while offering no solution whatsoever to the fundamental problem of how the information to build that system arose in the first place. The evidence, when properly interpreted, points not to unguided creation but to the sorting of designed potential within a system that is fundamentally decaying.
A Fair Summary of the Research
The research team focused on the folA gene in E. coli, which produces the essential metabolic enzyme dihydrofolate reductase (DHFR). This enzyme is the target of the antibiotic trimethoprim. The researchers used CRISPR-Cas9 gene editing to create a combinatorially complete library of variants for a specific nine-nucleotide (three-amino-acid) region of the gene known to be involved in antibiotic resistance.
They then subjected this massive population of bacterial variants to a “sublethal dose” of trimethoprim and used deep sequencing to measure the relative “fitness” (i.e., survival and replication rate) of each of the 261,382 variants for which they obtained data.
Their key findings were twofold:
- The landscape is rugged: They identified 514 distinct “fitness peaks,” or genotypes that were fitter than all of their immediate one-mutation neighbors. Classical theory suggests such ruggedness should trap evolving populations on low- or intermediate-fitness peaks, preventing them from reaching the optimal solution.
- The highest peaks are accessible: Despite the ruggedness, they found that the 74 highest-fitness peaks had enormous “basins of attraction.” This means that a large majority of all possible genotypes (over 90% in some cases) could reach a high-fitness peak through a series of short, exclusively beneficial (fitness-increasing) mutational steps. Their simulations of “adaptive walks” confirmed that 76.5% of populations starting from a random point would successfully reach one of these highly-fit peaks.
The authors conclude that landscape ruggedness is not the barrier to evolution it was once thought to be. Instead, it makes the evolutionary outcome highly “contingent,” or dependent on chance, because many different peaks are accessible from the same starting point.
The Core Critique
While the experimental work is impressive, its use as evidence for molecules-to-man evolution is invalidated by several fundamental flaws in logic. The study does not demonstrate the generation of new information, but rather the degradation of a pre-existing system.
The “Assume a Gene” Fallacy
The entire experiment begins with a fully-formed, functional, and information-rich biological system: a living E. coli bacterium with a working folA gene that codes for the complex DHFR enzyme. The study only explores the effects of minor tinkering within a tiny, three-amino-acid segment of this pre-existing machine. It says nothing about the origin of the DHFR protein fold, the genetic information in the folA gene, or the complex transcription and translation machinery required to produce it. This is a textbook case of what is sometimes called the “displacement problem”: explaining the modification of existing information, not its origin. This is not macroevolution.
The First Rule of Adaptive Evolution: Break Things to Survive
The “fitness” being measured in this experiment is not an improvement in the enzyme’s primary metabolic function. It is survival in the presence of a poison (trimethoprim). The easiest and fastest way to gain resistance to a targeted drug is to mutate the target enzyme in a way that hinders the drug’s ability to bind to it. This almost invariably involves a change that degrades or blunts the enzyme’s precise, original function.
This is a perfect illustration of biochemist Michael Behe’s “First Rule of Adaptive Evolution”: Break or blunt any functional gene whose loss would increase the number of a species’s offspring. The “navigable paths” to “high fitness” are not constructive evolutionary pathways; they are various routes to adaptive degeneration. They are different ways of breaking the lock so the poison’s key no longer fits. The fact that the authors found that an astonishing 93% of all variants were non-functional underscores this point. The functional space is an infinitesimal island in a vast sea of non-function, and the paths being explored are simply those that lead to the edge of the island, where the enzyme is just broken enough to evade the antibiotic but not so broken that the cell dies.
The Problem of Vanishingly Small Functional Space
Even within this tiny, 9-nucleotide space, the vast majority of sequences are useless. This provides a striking microcosm of the combinatorial inflation problem articulated by biochemist Douglas Axe. His research showed that for a modest 150-amino-acid protein, the ratio of functional sequences to non-functional ones is a staggering 1 in 10^77. If a mere three-amino-acid tweak renders DHFR non-functional 93% of the time, it stretches credulity to believe that a random search could ever build the entire enzyme from scratch. The “easy navigation” the authors found is only possible because their search was confined to the immediate vicinity of a pre-existing, highly-specified functional peak.
The Better Explanation: Designed Potential and Genetic Entropy
A far more robust explanation for these findings comes from a model of intelligent design, which posits that organisms were created with pre-existing potential for adaptation.
Designed for Adaptation
The E. coli genome is not a static blueprint that is slowly improved by random errors. It is a dynamic, responsive operating system. The “navigable paths” to antibiotic resistance are better understood not as lucky chance discoveries, but as the activation of a pre-programmed adaptive capacity. An intelligent engineer, knowing that the organism would face environmental challenges like toxins, would build in robust systems and potential pathways for adaptation. The navigability of the landscape is a design feature, not a product of chance. Selection, in this model, is not a creative force but a process that sorts through and activates this designed, front-loaded potential.
The Inescapable Reality of Genetic Entropy
This experiment provides a snapshot of short-term adaptation, but the universal, long-term trajectory of all complex genomes is one of decay. The Second Law of Thermodynamics, when applied to information systems, dictates that they will always lose information over time unless maintained by an intelligent agent. Genomes are relentlessly accumulating nearly-neutral mutations—typographical errors—that are too subtle for natural selection to see and remove.
While the bacteria in this study found a short-term fix, the overall process is one of degradation. The high number of non-functional variants (93%) is a direct, experimental confirmation of this entropic principle. The population is not climbing toward a new, higher form of life; it is simply finding clever ways to fall apart that happen to confer a temporary survival advantage in a poisoned environment. This is not the engine of evolution; it is the signature of decay.
Conclusion
Papkou and colleagues have performed a technologically remarkable experiment in operational science, mapping molecular cause-and-effect with incredible precision. However, extrapolating these findings to support the grand narrative of unguided, molecules-to-man evolution is a profound methodological error.
The study does not show the creation of new biological information. It shows the degradation of a complex, pre-existing system to achieve a short-term gain. It begins by assuming the existence of the very thing it needs to explain: a functionally-specified gene and its protein product. The “easily navigable” paths are not routes to genuinely new function, but pathways of adaptive degeneration, brilliantly illustrating how breaking things can sometimes be beneficial for survival.
When viewed through a more rigorous scientific lens, the evidence powerfully supports a model of designed life. Organisms appear to have been engineered with a robust, front-loaded capacity for adaptation, allowing them to respond to environmental challenges. The “navigability” of the fitness landscape is a testament to the foresight of that design. Yet, this adaptive potential exists within an overarching framework of genetic entropy, where the ultimate trajectory is one of decay, not the generation of the new information and machinery required to build a bacterium in the first place.
Leave a Reply