arXiv subjects

Piero Fariselli

Publications and source records attributed to Piero Fariselli.

15 recordsLinked to original sources

arXiv2026

Human mutation field reveals an equilibrium-like structure with irreversible circulation

The evolution of DNA sequences can be viewed as stochastic dynamics on a high-dimensional discrete space, but it is unclear when empirical transition biases reduce to an effective energy landscape versus retain irreducible non-equilibrium circulation. Human context-dependent mutation probabilities offer a direct test: every single-nucleotide substitution in a local context has a reverse substitution, so the logarithm of the forward-to-reverse probability ratio defines an antisymmetric field-the human mutation field. We show this field has a dominant gradient component and a smaller but reproducible curl component. Using seven-base human germline substitution probabilities, we infer an effective mutational landscape with a Siamese neural network constrained to predict only energy differences. This model predicts forward-to-reverse log-ratios for held-out mutations with a correlation of about 0.93, close to both an unconstrained predictive reference (0.948) and the empirical reversible ceiling from Hodge projection (about 0.96). Although trained only on mutation probabilities, the inferred landscape largely recovers short-word genomic composition and Chargaff reverse-complement symmetry for sequences up to length four. Deviations from equilibrium structure reveal a small but detectable nonequilibrium component: a residual irreversible circulation violating the Kolmogorov cycle condition for detailed balance, reproducible across African, Asian, and European populations, and strongest in CpG-linked cycles and CpG-transition edges, consistent with methylcytosine deamination. These results give a thermodynamic decomposition of the human mutation field: most mutation bias is organized by a local equilibrium-like energy landscape aligned with genome composition, while the residual circulation points to specific directional mutational mechanisms.

q-bio.GN↗

arXiv2025

Neural Diffusion Processes for Physically Interpretable Survival Prediction

We introduce DeepFHT, a survival-analysis framework that couples deep neural networks with first hitting time (FHT) distributions from stochastic process theory. Time to event is represented as the first passage of a latent diffusion process to an absorbing boundary. A neural network maps input variables to physically meaningful parameters including initial condition, drift, and diffusion, within a chosen FHT process such as Brownian motion, both with drift and driftless. This yields closed- form survival and hazard functions and captures time-varying risk without assuming proportional- hazards. We compare DeepFHT with Cox regression using synthetic and real-world datasets. The method achieves predictive accuracy on par with the state-of-the-art approach, while maintaining a physics- based interpretable parameterization that elucidates the relation between input features and risk. This combination of stochastic process theory and deep learning provides a principled avenue for modeling survival phenomena in complex systems

cs.LG↗

arXiv2025

The Advantage of Fine-Grained Training

In classification problems, models are trained to predict a class label based on the input data features. However, class labels are organized hierarchically in many datasets. While a classification task is often defined at a specific level of this hierarchy, training can utilize a finer granularity of labels. Empirical evidence suggests that such fine-grained training can enhance performance. In this work, we investigate the generality of this observation and explore its underlying causes using both real and synthetic datasets. We show that training on fine-grained labels does not universally improve classification accuracy. Instead, the effectiveness of this strategy depends on the geometric structure of the data and its relations with the label hierarchy. Specifically, we show that the advantage of fine-grained training crucially depends on the degree of alignment between the decision boundaries required for the fine- and coarse-grained tasks, a property that we term boundary redundancy. Additionally, factors such as dataset size and model capacity significantly influence whether fine-grained labels provide a performance benefit. Indeed, we identify a transition, whose location is largely controlled by the degree of overparameterization, separating regimes where fine-grained training improves performance from those where direct coarse-grained training is preferable.

cs.LG↗

arXiv2025

Sparsity is All You Need: Rethinking Biological Pathway-Informed Approaches in Deep Learning

Biologically-informed neural networks typically leverage pathway annotations to enhance performance in biomedical applications. We hypothesized that the benefits of pathway integration does not arise from its biological relevance, but rather from the sparsity it introduces. We conducted a comprehensive analysis of all relevant pathway-based neural network models for predictive tasks, critically evaluating each study's contributions. From this review, we curated a subset of methods for which the source code was publicly available. The comparison of the biologically informed state-of-the-art deep learning models and their randomized counterparts showed that models based on randomized information performed equally well as biologically informed ones across different metrics and datasets. Notably, in 3 out of the 15 analyzed models, the randomized versions even outperformed their biologically informed counterparts. Moreover, pathway-informed models did not show any clear advantage in interpretability, as randomized models were still able to identify relevant disease biomarkers despite lacking explicit pathway information. Our findings suggest that pathway annotations may be too noisy or inadequately explored by current methods. Therefore, we propose a methodology that can be applied to different domains and can serve as a robust benchmark for systematically comparing novel pathway-informed models against their randomized counterparts. This approach enables researchers to rigorously determine whether observed performance improvements can be attributed to biological insights.

q-bio.QM↗

arXiv2025

Beyond Cox Models: Assessing the Performance of Machine-Learning Methods in Non-Proportional Hazards and Non-Linear Survival Analysis

Survival analysis often relies on Cox models, assuming both linearity and proportional hazards (PH). This study evaluates machine and deep learning methods that relax these constraints, comparing their performance with penalized Cox models on a benchmark of three synthetic and three real datasets. In total, eight different models were tested, including six non-linear models of which four were also non-PH. Although Cox regression often yielded satisfactory performance, we showed the conditions under which machine and deep learning models can perform better. Indeed, the performance of these methods has often been underestimated due to the improper use of Harrell's concordance index (C-index) instead of more appropriate scores such as Antolini's concordance index, which generalizes C-index in cases where the PH assumption does not hold. In addition, since occasionally high C-index models happen to be badly calibrated, combining Antolini's C-index with Brier's score is useful to assess the overall performance of a survival method. Results on our benchmark data showed that survival prediction should be approached by testing different methods to select the most appropriate one according to sample size, non-linearity and non-PH conditions. To allow an easy reproducibility of these tests on our benchmark data, code and documentation are freely available at https://github.com/compbiomed-unito/survhive.

cs.LG↗

arXiv2025

Mass Balance Approximation of Unfolding Improves Potential-Like Methods for Protein Stability Predictions

The prediction of protein stability changes following single-point mutations plays a pivotal role in computational biology, particularly in areas like drug discovery, enzyme reengineering, and genetic disease analysis. Although deep-learning strategies have pushed the field forward, their use in standard workflows remains limited due to resource demands. Conversely, potential-like methods are fast, intuitive, and efficient. Yet, these typically estimate Gibbs free energy shifts without considering the free-energy variations in the unfolded protein state, an omission that may breach mass balance and diminish accuracy. This study shows that incorporating a mass-balance correction (MBC) to account for the unfolded state significantly enhances these methods. While many machine learning models partially model this balance, our analysis suggests that a refined representation of the unfolded state may improve the predictive performance.

q-bio.QM↗

arXiv2025

JanusDDG: A Thermodynamics-Compliant Model for Sequence-Based Protein Stability via Two-Fronts Multi-Head Attention

Understanding how residue variations affect protein stability is crucial for designing functional proteins and deciphering the molecular mechanisms underlying disease-related mutations. Recent advances in protein language models (PLMs) have revolutionized computational protein analysis, enabling, among other things, more accurate predictions of mutational effects. In this work, we introduce JanusDDG, a deep learning framework that leverages PLM-derived embeddings and a bidirectional cross-attention transformer architecture to predict $\Delta \Delta G$ of single and multiple-residue mutations while simultaneously being constrained to respect fundamental thermodynamic properties, such as antisymmetry and transitivity. Unlike conventional self-attention, JanusDDG computes queries (Q) and values (V) as the difference between wild-type and mutant embeddings, while keys (K) alternate between the two. This cross-interleaved attention mechanism enables the model to capture mutation-induced perturbations while preserving essential contextual information. Experimental results show that JanusDDG achieves state-of-the-art performance in predicting $\Delta \Delta G$ from sequence alone, matching or exceeding the accuracy of structure-based methods for both single and multiple mutations. Code Availability:https://github.com/compbiomed-unito/JanusDDG

q-bio.QM↗

arXiv2025

SurvHive: a package to consistently access multiple survival-analysis packages

Survival analysis, a foundational tool for modeling time-to-event data, has seen growing integration with machine learning (ML) approaches to handle the complexities of censored data and time-varying risks. Despite these advances, leveraging state-of-the-art survival models remains a challenge due to the fragmented nature of existing implementations, which lack standardized interfaces and require extensive preprocessing. We introduce SurvHive, a Python-based framework designed to unify survival analysis methods within a coherent and extensible interface modeled on scikit-learn. SurvHive integrates classical statistical models with cutting-edge deep learning approaches, including transformer-based architectures and parametric survival models. Using a consistent API, SurvHive simplifies model training, evaluation, and optimization, significantly reducing the barrier to entry for ML practitioners exploring survival analysis. The package includes enhanced support for hyper-parameter tuning, time-dependent risk evaluation metrics, and cross-validation strategies tailored to censored data. With its extensibility and focus on usability, SurvHive provides a bridge between survival analysis and the broader ML community, facilitating advancements in time-to-event modeling across domains. The SurvHive code and documentation are available freely at https://github.com/compbiomed-unito/survhive.

q-bio.QM↗

arXiv2023

How good Neural Networks interpretation methods really are? A quantitative benchmark

Saliency Maps (SMs) have been extensively used to interpret deep learning models decision by highlighting the features deemed relevant by the model. They are used on highly nonlinear problems, where linear feature selection (FS) methods fail at highlighting relevant explanatory variables. However, the reliability of gradient-based feature attribution methods such as SM has mostly been only qualitatively (visually) assessed, and quantitative benchmarks are currently missing, partially due to the lack of a definite ground truth on image data. Concerned about the apophenic biases introduced by visual assessment of these methods, in this paper we propose a synthetic quantitative benchmark for Neural Networks (NNs) interpretation methods. For this purpose, we built synthetic datasets with nonlinearly separable classes and increasing number of decoy (random) features, illustrating the challenge of FS in high-dimensional settings. We also compare these methods to conventional approaches such as mRMR or Random Forests. Our results show that our simple synthetic datasets are sufficient to challenge most of the benchmarked methods. TreeShap, mRMR and LassoNet are the best performing FS methods. We also show that, when quantifying the relevance of a few non linearly-entangled predictive features diluted in a large number of irrelevant noisy variables, neural network-based FS and interpretation methods are still far from being reliable.

cs.LG↗

arXiv2019

DNA energy constraints shape biological evolutionary trajectories

Most living systems rely on double-stranded DNA (dsDNA) to store their genetic information and perpetuate themselves. This biological information has been considered the main target of evolution. However, here we show that symmetries and patterns in the dsDNA sequence can emerge from the physical peculiarities of the dsDNA molecule itself and the maximum entropy principle alone, rather than from biological or environmental evolutionary pressure. The randomness justifies the human codon biases and context-dependent mutation patterns in human populations. Thus, the DNA "exceptional symmetries", emerged from the randomness, have to be taken into account when looking for the DNA encoded information. Our results suggest that the double helix energy constraints and, more generally, the physical properties of the dsDNA are the hard drivers of the overall DNA sequence architecture, whereas the biological selective processes act as soft drivers, which only under extraordinary circumstances overtake the overall entropy content of the genome.

q-bio.BM↗

arXiv2018

A natural upper bound to the accuracy of predicting protein stability changes upon mutations

Accurate prediction of protein stability changes upon single-site variations (DDG) is important for protein design, as well as our understanding of the mechanism of genetic diseases. The performance of high-throughput computational methods to this end is evaluated mostly based on the Pearson correlation coefficient between predicted and observed data, assuming that the upper bound would be 1 (perfect correlation). However, the performance of these predictors can be limited by the distribution and noise of the experimental data. Here we estimate, for the first time, a theoretical upper-bound to the DDG prediction performances imposed by the intrinsic structure of currently available DDG data. Given a set of measured DDG protein variations, the theoretically best predictor is estimated based on its similarity to another set of experimentally determined DDG values. We investigate the correlation between pairs of measured DDG variations, where one is used as a predictor for the other. We analytically derive an upper bound to the Pearson correlation as a function of the noise and distribution of the DDG data. We also evaluate the available datasets to highlight the effect of the noise in conjunction with DDG distribution. We conclude that the upper bound is a function of both uncertainty and spread of the DDG values, and that with current data the best performance should be between 0.7-0.8, depending on the dataset used; higher Pearson correlations might be indicative of overtraining. It also follows that comparisons of predictors using different datasets are inherently misleading.

q-bio.BM↗

arXiv2015

Network measures for protein folding state discrimination

Proteins fold using a two-state or multi-state kinetic mechanisms, but up to now there isn't a first-principle model to explain this different behaviour. We exploit the network properties of protein structures by introducing novel observables to address the problem of classifying the different types of folding kinetics. These observables display a plain physical meaning, in terms of vibrational modes, possible configurations compatible with the native protein structure, and folding cooperativity. The relevance of these observables is supported by a classification performance up to 90%, even with simple classifiers such as disciminant analysis.

q-bio.MN↗

arXiv2007

A three-state prediction of single point mutations on protein stability changes

A basic question of protein structural studies is to which extent mutations affect the stability. This question may be addressed starting from sequence and/or from structure. In proteomics and genomics studies prediction of protein stability free energy change (DDG) upon single point mutation may also help the annotation process. The experimental SSG values are affected by uncertainty as measured by standard deviations. Most of the DDG values are nearly zero (about 32% of the DDG data set ranges from -0.5 to 0.5 Kcal/mol) and both the value and sign of DDG may be either positive or negative for the same mutation blurring the relationship among mutations and expected DDG value. In order to overcome this problem we describe a new predictor that discriminates between 3 mutation classes: destabilizing mutations (DDG<-0.5 Kcal/mol), stabilizing mutations (DDG>0.5 Kcal/mol) and neutral mutations (-0.5<=DDG<=0.5 Kcal/mol). In this paper a support vector machine starting from the protein sequence or structure discriminates between stabilizing, destabilizing and neutral mutations. We rank all the possible substitutions according to a three state classification system and show that the overall accuracy of our predictor is as high as 52% when performed starting from sequence information and 58% when the protein structure is available, with a mean value correlation coefficient of 0.30 and 0.39, respectively. These values are about 20 points per cent higher than those of a random predictor.

q-bio.BM↗

arXiv2007

In silico evidence of the relationship between miRNAs and siRNAs

Both short interfering RNAs (siRNAs) and microRNAs (miRNAs) mediate the repression of specific sequences of mRNA through the RNA interference pathway. In the last years several experiments have supported the hypothesis that siRNAs and miRNAs may be functionally interchangeable, at least in cultured cells. In this work we verify that this hypothesis is also supported by a computational evidence. We show that a method specifically trained to predict the activity of the exogenous siRNAs assigns a high silencing level to experimentally determined human miRNAs. This result not only supports the idea of siRNAs and miRNAs equivalence but indicates that it is possible to use computational tools developed using synthetic small interference RNAs to investigate endogenous miRNAs.

q-bio.BM↗

arXiv2005

The posterior-Viterbi: a new decoding algorithm for hidden Markov models

Background: Hidden Markov models (HMM) are powerful machine learning tools successfully applied to problems of computational Molecular Biology. In a predictive task, the HMM is endowed with a decoding algorithm in order to assign the most probable state path, and in turn the class labeling, to an unknown sequence. The Viterbi and the posterior decoding algorithms are the most common. The former is very efficient when one path dominates, while the latter, even though does not guarantee to preserve the automaton grammar, is more effective when several concurring paths have similar probabilities. A third good alternative is 1-best, which was shown to perform equal or better than Viterbi. Results: In this paper we introduce the posterior-Viterbi (PV) a new decoding which combines the posterior and Viterbi algorithms. PV is a two step process: first the posterior probability of each state is computed and then the best posterior allowed path through the model is evaluated by a Viterbi algorithm. Conclusions: We show that PV decoding performs better than other algorithms first on toy models and then on the computational biological problem of the prediction of the topology of beta-barrel membrane proteins.

q-bio.BM↗