arXiv subjects

Vincenzo Carnevale

Publications and source records attributed to Vincenzo Carnevale.

9 recordsLinked to original sources

arXiv2026

HyperEvoGen: Exploring deep phylogeny using non-Euclidean variational inference

Homologous proteins evolve from a common ancestral sequence, constrained by intricate patterns of co-evolving residues. Accurate reconstruction of evolutionary histories remains a challenge, primarily due to the inability of the existing approaches to capture long-range coevolutionary ties and lack of a precise metric to represent the evolutionary distance between sequences. Standard approaches are based on p-distance or substitution-corrected measures such as Jukes-Cantor. These methods saturate in cases of deep evolutionary divergence, losing all evolutionary signal after enough time. We present HyperEvoGen, a Poincaré variational autoencoder with adversarial training, hyperbolic latent geometry, and a compound loss function that learns evolutionarily meaningful representations from single-family alignments. The arrangement of protein sequences in HyperEvoGen's hyperbolic embedding aims to preserve phylogenetic structure and produce latent distances which scale with true evolutionary divergence. HyperEvoGen enables fast, scalable modeling of protein evolution while preserving hierarchical relatedness in a geometry-aware representation. On Potts-coupled simulation benchmarks, it produces more accurate ancestral reconstructions than conventional baselines, and offers higher-quality sequence generation with less training time than Potts models. This combination of accuracy and throughput supports large-family evolutionary studies and accelerates design-oriented applications.

q-bio.QM↗

arXiv2025

Generative Modeling of Entangled Polymers with a Distance-Based Variational Autoencoder

We present a variational autoencoder framework for learning and generating configurations of structured polymer globules from distance matrices. We used coarse-grained molecular dynamics to sample polyethylene structures, which we used as the training set for our deep learning model. By combining convolution and attention layers, the model encodes the structural patterns of distance matrices into an organized and roto-translationally invariant latent space of lower dimensionality. The generative capability of the variational autoencoder, coupled with a post-processing pipeline based on multidimensional scaling and short molecular dynamics, enables the recovery of physically meaningful configurations. The reconstructed and generated samples reproduce key observables, including energy, size, and entanglement, despite minor discrepancies in the raw decoder output.

physics.comp-ph↗

arXiv2024

Lipid-mediated hydrophobic gating in the BK potassium channel

The large-conductance, calcium-activated potassium (BK) channel lacks the typical intracellular bundle-crossing gate present in most ion channels of the 6TM family. This observation, initially inferred from Ca$^{2+}$-free-pore accessibility experiments and recently corroborated by a CryoEM structure of the non-conductive state, raises a puzzling question: how can gating occur in absence of steric hindrance? To answer this question, we carried out molecular simulations and accurate free energy calculations to obtain a microscopic picture of the sequence of events that, starting from a Ca$^{2+}$-free state leads to ion conduction upon Ca$^{2+}$ binding. Our results highlight an unexpected role for annular lipids, which turn out to be an integral part of the gating machinery. Due to the presence of fenestrations, the "closed" Ca$^{2+}$-free pore can be occupied by the methyl groups from the lipid alkyl chains. This dynamic occupancy triggers and stabilizes the nucleation of a vapor bubble into the inner pore cavity, thus hindering ion conduction. By contrast, Ca$^{2+}$ binding results into a displacement of these lipids outside the inner cavity, lowering the hydrophobicity of this region and thus allowing for pore hydration and conduction. This lipid-mediated hydrophobic gating rationalizes several seemingly problematic experimental observations, including the state-dependent pore accessibility of blockers.

q-bio.BM↗

arXiv2024

Ion channels in critical membranes: clustering, cooperativity, and memory effects

Much progress has been made in elucidating the inner workings of voltage-gated ion channels, but less understood is the influence of lipid rafts on gating kinetics. Here we propose that state-dependent channel affinity for different lipid species provides a unified explanation for the experimentally observed behaviors of clustering, cooperativity, and hysteresis. We develop models of diffusing lipids and channels engaged in Ising-like interactions to investigate the collective behaviors driven by raft formation in critical membranes close to the demixing transition. The model channels demonstrate lipid-mediated long-range interactions, activation curve steepening, and long-term memory in ionic currents. These behaviors likely play a role in channel-mediated cellular signaling and suggest a universal mechanism for self-organization of biomolecular assemblies.

cond-mat.soft↗

arXiv2022

Nonequilibrium thermodynamics of DNA nanopore unzipping

Using theory and simulations, we carried out a first systematic characterization of DNA unzipping via nanopore translocation. Starting from partially unzipped states, we found three dynamical regimes depending on the applied force, f: (i) heterogeneous DNA retraction and rezipping (f < 17pN), (ii) normal (17pN < f < 60pN) and (iii) anomalous (f > 60pN) drift-diffusive behavior. We show that the normal drift-diffusion regime can be effectively modelled as a one-dimensional stochastic process in a tilted periodic potential. We use the theory of stochastic processes to recover the potential from nonequilibrium unzipping trajectories and show that it corresponds to the free-energy landscape for single base-pairs unzipping. Applying this general approach to other single-molecule systems with periodic potentials ought to yield detailed free-energy landscapes from out-of-equilibrium trajectories.

cond-mat.soft↗

arXiv2021

Generative Capacity of Probabilistic Protein Sequence Models

Potts models and variational autoencoders (VAEs) have recently gained popularity as generative protein sequence models (GPSMs) to explore fitness landscapes and predict the effect of mutations. Despite encouraging results, quantitative characterization and comparison of GPSM-generated probability distributions is still lacking. It is currently unclear whether GPSMs can faithfully reproduce the complex multi-residue mutation patterns observed in natural sequences arising due to epistasis. We develop a set of sequence statistics to assess the "generative capacity" of three GPSMs of recent interest: the pairwise Potts Hamiltonian, the VAE, and the site-independent model, using natural and synthetic datasets. We show that the generative capacity of the Potts Hamiltonian model is the largest, in that the higher order mutational statistics generated by the model agree with those observed for natural sequences. In contrast, we show that the VAE's generative capacity lies between the pairwise Potts and site-independent models. Importantly, our work measures GPSM generative capacity in terms of higher-order sequence covariation statistics which we have developed, and provides a new framework for evaluating and interpreting GPSM accuracy that emphasizes the role of epistasis.

cs.LG↗

arXiv2020

A particle-based Ising model

We characterize equilibrium properties and relaxation dynamics of a two-dimensional lattice containing, at each site, two particles connected by a double-well potential (dumbbell). Dumbbells are oriented in the orthogonal direction with respect to the lattice plane and interact with each other through a Lennard-Jones potential truncated at the nearest neighbor distance. We show that the system's equilibrium properties are accurately described by a two-dimensional Ising model with an appropriate coupling constant. Moreover, we characterize the coarsening kinetics by calculating the cluster size as a function of time and compare the results with Monte Carlo simulations based on Glauber or reactive dynamics rate constants.

cond-mat.stat-mech↗

arXiv2018

Is the free energy landscape informative about transition rates? Lessons from the kinetic Ising model

An oft-used concept in modeling macromolecules is the free energy landscape, obtained by coarse-graining a vast number of microstates into a low-dimensional mesh of mesostates. If the landscape contains two or more local minima (macrostates),one can compute global rate constants provided the dynamics of the dividing barrier regions are known. Here we compared experimental rate constants between ordered states in a kinetic Ising model with rates calculated from a coarse-grained master equation derived from the microcanonical ensemble. The coarse-grained macroscopic rate constants were roughly 50 % larger than experiment across a range of environmental constraints, suggesting a systematic impediment of configurational progress on the microscopic scale that is specific to the structure of the Ising model. The error in coarse-graining lay with the calculation of the diffusion coefficient rather than with the shape of the free energy landscape, as ensemble- and time-averaged estimates of the latter were indistinguishable. Fluctuation analysis in the form of Nyquist theorem also failed to substantially improve the value of the effective diffusion coefficient, suggesting a failure of the fluctuation-dissipation theorem. These findings from the Ising model raises doubts over the validity of the free energy landscape approach in calculating absolute transition rates for more complex systems such as proteins.

cond-mat.stat-mech↗

arXiv2006

Convergent dynamics in the protease enzymatic superfamily

Proteases regulate various aspects of the life cycle in all organisms by cleaving specific peptide bonds. Their action is so central for biochemical processes that at least 2% of any known genome encodes for proteolytic enzymes. Here we show that selected proteases pairs, despite differences in oligomeric state, catalytic residues and fold, share a common structural organization of functionally relevant regions which are further shown to undergo similar concerted movements. The structural and dynamical similarities found pervasively across evolutionarily distant clans point to common mechanisms for peptide hydrolysis.

q-bio.BM↗