Nicolas Lartillot

The paper tests the nearly-neutral theory across 144 placental mammal genomes, showing correlations between genetic diversity, life-history traits, and selection efficacy at both micro- and macro-evolutionary scales.

The paper proposes a mathematical model showing how genome size evolution is shaped by mutation biases and robustness selection, leading to an equilibrium non-coding fraction.

This study shows that GC-biased gene conversion (gBGC) evolves under weak stabilising selection and can increase population load, suggesting it may have been positively selected in humans.

This study uses computational models to show that population size and mutation rate interact to determine genome size and coding density in bacteria.

The paper compares model selection methods in Bayesian phylogenetics, finding cross-validation and wAIC outperform Bayes factors for selecting the best-approximating model.

The paper introduces SDpop, a method to detect sex-linked genes in natural populations using genotyped individuals, showing robust predictions with 5-10 individuals per sex.

The paper introduces EDCluster, a scalable method for empirical distribution mixture models that reduces long branch attraction artifacts and improves performance over existing models.

This paper clarifies how convergent adaptive evolution manifests in coding sequences, reviews detection methods, and evaluates their power through simulations.