Daniele Silvestro

This study evaluates spike clock models in phylogenetics, finding they are supported in most datasets and improve estimates of speciation and extinction rates.

The paper reanalyzes neoselachian diversity through the K/Pg boundary using new data and methods, finding no diversity collapse during the K/Pg and robust results across different data treatments.

This paper develops a reinforcement learning framework for conservation planning that accounts for climate-driven range shifts and niche coverage, improving conservation priorities for European crop wild relatives.

This paper introduces BELLA, a Bayesian neural network framework for modeling complex phylodynamic processes, showing improved accuracy over linear models in capturing nonlinear relationships between predictors and evolutionary rates.

The paper identifies global conservation priorities for sharks and rays by analyzing threat levels, functional roles, and evolutionary distinctiveness, revealing protection deficits and conflicts with fishing pressures.

The paper models how immigration of African and North American mammals affected South American mammal diversity, finding increased overall richness but declining native diversity over time.

The paper analyzes 27 radiations using a new fossilized birth-death model to show that clade diversity follows rise and decline dynamics driven by species-level processes rather than clade-level mechanisms.

DeepDiveR estimates palaeodiversity from fossil occurrences using deep learning, accounting for preservation biases and showing a 37% species loss in Carnivora since the Pleistocene.

This paper uses multi-agent reinforcement learning to simulate animal behavior in alpine ecosystems and assesses ecosystem resilience under various stress scenarios.

This study uses simulations to show that environmental instability increases extinction rates and phenotypic evolution, affecting species with limited adaptability.

This paper uses AI to optimize ecological restoration by balancing biodiversity and carbon sequestration goals, showing that prioritizing biodiversity can enhance conservation while slightly reducing carbon gains.

The paper investigates selection relaxation in aquatic mammals and finds it occurs in 20% of genes across both semiaquatic and fully aquatic species.

The paper shows that estimating species age from phylogenetic trees is error-prone due to sampling biases and speciation assumptions, and proposes a probabilistic model to improve accuracy.

This study benchmarks imputation methods for discrete biological traits using a pipeline combining phylogenetic, machine learning, and deep learning approaches, finding an ensemble method to be most accurate.

raxmlGUI 2.0 provides a graphical interface for RAxML, enabling phylogenetic analyses with improved automation and support for new RAxML versions.

SpeciesGeoCoder categorizes species occurrences into user-defined areas for biodiversity and evolutionary analyses, enabling fast and flexible data processing.