Vol. 1, No. 35Friday, November 20, 2026Bryan, Texas
One paper worth the coffee. One paper from beyond the fence. One story from the wider world.
The lead · Epistasis and protein sequence evolution models
Gene interactions hide evolutionary distance
A continuous-time protein model shows how epistasis can slow sequence evolution without changing average rates at individual sites.
Pagnani A, Barrat-Charlaix P·Genome Biology and Evolution, 17 Nov 2026·doi:10.1093/gbe/evag281
Most evolutionary models treat each position in a protein as though its substitutions occur independently. That assumption misses epistasis: a mutation's effects depend on the amino acids at other positions. Andrea Pagnani and Pierre Barrat-Charlaix built a continuous-time simulation framework that allows positions to interact. Their model draws on a Potts model, a statistical description of which amino acid combinations tend to occur together, and simulates individual substitutions using the Gillespie algorithm. This lets researchers follow a protein sequence through evolutionary time while retaining the interdependencies learned from protein families.
Compared with models in which sites evolve independently, these simulations produced a counterintuitive result. Epistasis slowed evolutionary change overall, even though average substitution rates at individual positions were not altered. The apparent contradiction emerged because site-specific rates shifted with sequence context: some positions sometimes became effectively frozen and at other times changed rapidly. The team also found that standard methods could underestimate the evolutionary distance between sequences generated under epistasis. That matters for phylogenetics, where sequence differences are used to reconstruct evolutionary histories. The study demonstrates a modeling approach, not a new measurement of mutation rates across organisms; whether its biases strongly affect real phylogenies remains an empirical question.
Where to pushTest the effect on inference rather than simulation alone. If empirical protein alignments with strong residue coupling are analyzed under standard independent-site models and under models informed by those couplings, do branch lengths or divergence times shift consistently? Simulated benchmarks could separate error caused by epistasis from error caused by rate variation unrelated to interactions.
Beyond EEB · A paper from outside evolution and ecology
A glucose-powered implant closes a nerve-immune feedback loop
An experimental implant uses glucose inside the body both as its energy source and as a signal to regulate nerve activity. The researchers designed an artificial neuroimmune feedback system that mimics aspects of the vagal reflex, a pathway linking the nervous and immune systems. It requires neither a battery nor wireless power. In rat models of two different forms of diabetes, the system reduced inflammatory responses and improved measures of glucose regulation. Those effects reportedly persisted after treatment stopped. The result is a demonstration in animals, not a validated human therapy, but it suggests a route toward self-powered devices that respond to changing physiology.
History · From the history of science
Lynn Margulis made cooperation central to the origin of complex cells
Dolan, Journal of Eukaryotic Microbiology (2012), in memoriam.
In 1967, Lynn Margulis published a synthesis arguing that major components of complex cells arose when formerly independent microorganisms entered long-lasting partnerships. The idea was initially resisted, although evidence for symbiotic origins had been proposed by earlier scientists. Over time, molecular and cellular evidence firmly established the bacterial ancestry of mitochondria and chloroplasts, helping transform how biologists explain the origin of eukaryotes. Margulis, who died in 2011, remained an energetic advocate of symbiosis as an evolutionary force. She also advanced broader claims, including a bacterial origin for eukaryotic flagella, that did not gain comparable evidential support. Her story illustrates both the value of revisiting neglected hypotheses and the need to test each proposed evolutionary merger separately.