In recent years, generative models have shown remarkable capabilities in synthesizing realistic human motion, with applications ranging from animation to virtual reality. However, their potential in clinical and rehabilitation settings remains underexplored. In this work, we introduce a conditional diffusion-based generative framework for rehabilitation-oriented motion synthesis, which directly operates on joint-angle representations of full-body movement. Unlike most existing approaches that rely on joint positions, our method generates motion in a clinically meaningful space that explicitly encodes joint range of motion, aligning the generation process with how motor performance is assessed in rehabilitation practice. This design enables subject-independent modeling while improving the interpretability of the generated movements from a clinical perspective. We propose a comprehensive evaluation protocol by combining qualitative and quantitative metrics, including simulation visualizations, similarity analysis, and automated assessment of simulations adherence to users input. Experiments based on cross-subject and leave-one-combination-out settings demonstrate the model's ability to generate plausible, contextually accurate motion sequences, with improved generalization when using joint angle representations, achieving superior performance compared to a position-based approach. Despite limitations due to dataset size and gesture diversity, results support the feasibility of generating rehabilitation-oriented motion simulations, motivating future investigation in personalized rehabilitation scenarios.

BACKGROUND: The COVID-19 pandemic has caused over 776 million infections and 7 million deaths globally between December 2019 and November 2024. Since the emergence of the original Wuhan strain, SARS-CoV-2 has evolved into multiple variants-including Alpha, Delta, and Omicron-primarily through mutations in the Spike glycoprotein. The S1 subunit, which binds the human angiotensin-converting enzyme 2 (ACE2) receptor, mutates frequently and plays a key role in infectivity and immune escape, while the more conserved S2 subunit mediates membrane fusion. Anticipating future mutations is essential for guiding vaccine design and therapeutic strategies. Generative Large Language Models (LLMs) have shown promise in protein sequence modeling due to their capacity to produce realistic and functional synthetic sequences. Here, we introduce SARITA, a GPT-3-based LLM with up to 1.2 billion parameters, fine-tuned via continual learning on the protein model RITA trained on 107 017 high-quality SARS-CoV-2 Spike sequences (up to March 1st 2021) to generate high-quality synthetic SARS-CoV-2 Spike S1 subunits. RESULTS: SARITA is able to generate realistic, full-length synthetic S1 subunits starting from a 14-amino-acid prompt. When evaluated on unseen sequences collected between March 2021 and November 2023-including major Variants of Concern (VOCs) such as Delta and Omicron, and Variants of Interest such as Iota-SARITA outperforms baseline and state-of-the-art LLMs in terms of sequence quality, biological plausibility, and similarity to real-world viral evolution. SARITA generates high-quality sequences in over 97% of cases, with markedly lower False Mutation Rate and higher similarity scores (PAM30, Levenshtein distance) compared to alternative approaches. It also accurately reproduces key mutations characteristic of future variants-such as L212I, R158L, T95P, and E406K-which were not present in the training data but emerged later in VOCs like Omicron and Delta. Structure-based analysis confirms the functional plausibility of these substitutions, with ΔΔG values within experimentally supported thresholds for ACE2 and antibody binding. Furthermore, SARITA anticipates immune-evasive mutations and accurately captures the positional and statistical distribution of mutations found in post- March 1st 2021 variants, highlighting its potential as a predictive tool for viral evolution. CONCLUSION: These results indicate the potential of SARITA to predict future SARS-CoV-2 S1 evolution, potentially aiding in the development of adaptable vaccines and treatments.

Applications of Artificial Intelligence (AI) are revolutionizing biomedical research and healthcare by offering data-driven predictions that assist in diagnoses. Supervised learning systems are trained on large datasets to predict outcomes for new test cases. However, they typically do not provide an indication of the reliability of these predictions, even though error estimates are integral to model development. Here, we introduce a novel method to identify regions in the feature space that diverge from training data, where an AI model may perform poorly. We utilize a compact precompiled structure that allows for fast and direct access to confidence scores in real time at the point of use without requiring access to the training data or model algorithms. As a result, users can determine when to trust the AI model’s outputs, while developers can identify where the model’s applicability is limited. We validate our approach using simulated data and several biomedical case studies, demonstrating that our approach provides fast confidence estimates ([Formula: see text] milliseconds per case), with high concordance to previously developed methods (f-[Formula: see text]). These estimates can be easily added to real-world AI applications. We argue that providing confidence estimates should be a standard practice for all AI applications in public use.

The COVID-19 pandemic has profoundly impacted global health, leading to millions of deaths and overwhelming healthcare systems worldwide. This study investigates the relationship between SARS-CoV-2 sequencing rates and critical epidemiological parameters, such as cases, deaths, and ICU admissions, across 25 European countries from January 2020 to November 2023. By analyzing these relationships, we aim to determine whether sequencing efforts were reactive—in response to epidemiological pressures—or proactive, guided by public health strategies. The analysis used publicly available data from GISAID, OxCGRT, and ECDC, and included weekly aggregation, correlation analysis, and the application of TimeGPT for predictive modeling. Results show that sequencing rates were significantly correlated with ICU admissions, hospitalizations, case numbers, and deaths, though with variability between countries and over different pandemic phases. TimeGPT analysis revealed that sequencing rates were often the most informative feature for predicting future COVID-19 cases in many countries. These findings highlight the potential of sequencing rates to serve as early indicators for severe pandemic outcomes and underscore the importance of context-specific approaches for managing future health crises.

The COVID-19 pandemic is marked by the successive emergence of new SARS-CoV-2 variants, lineages, and sublineages that outcompete earlier strains, largely due to factors like increased transmissibility and immune escape. We propose DeepAutoCoV, an unsupervised deep learning anomaly detection system, to predict future dominant lineages (FDLs). We define FDLs as viral (sub)lineages that will constitute >10% of all the viral sequences added to the GISAID, a public database supporting viral genetic sequence sharing, in a given week. DeepAutoCoV is trained and validated by assembling global and country-specific data sets from over 16 million Spike protein sequences sampled over a period of ~4 years. DeepAutoCoV successfully flags FDLs at very low frequencies (0.01%-3%), with median lead times of 4-17 weeks, and predicts FDLs between ~5 and ~25 times better than a baseline approach. For example, the B.1.617.2 vaccine reference strain was flagged as FDL when its frequency was only 0.01%, more than a year before it was considered for an updated COVID-19 vaccine. Furthermore, DeepAutoCoV outputs interpretable results by pinpointing specific mutations potentially linked to increased fitness and may provide significant insights for the optimization of public health 'pre-emptive' intervention strategies.

Increasingly complex learning methods such as boosting, bagging and deep learning have made ML models more accurate, but harder to interpret and explain, culminating in black-box machine learning models. Model developers and users alike are often presented with a trade-off between performance and intelligibility, especially in high-stakes applications like medicine. In the present article we propose a novel methodological approach for generating explanations for the predictions of a generic machine learning model, given a specific instance for which the prediction has been made. The method, named AraucanaXAI, is based on surrogate, locally-fitted classification and regression trees that are used to provide post-hoc explanations of the prediction of a generic machine learning model. Advantages of the proposed XAI approach include superior fidelity to the original model, ability to deal with non-linear decision boundaries, and native support to both classification and regression problems. We provide a packaged, open-source implementation of the AraucanaXAI method and evaluate its behaviour in a number of different settings that are commonly encountered in medical applications of AI. These include potential disagreement between the model prediction and physician's expert opinion and low reliability of the prediction due to data scarcity.

Interest in Machine Learning applications to tackle clinical and biological problems is increasing. This is driven by promising results reported in many research papers, the increasing number of AI-based software products, and by the general interest in Artificial Intelligence to solve complex problems. It is therefore of importance to improve the quality of machine learning output and add safeguards to support their adoption. In addition to regulatory and logistical strategies, a crucial aspect is to detect when a Machine Learning model is not able to generalize to new unseen instances, which may originate from a population distant to that of the training population or from an under-represented subpopulation. As a result, the prediction of the machine learning model for these instances may be often wrong, given that the model is applied outside its "reliable" space of work, leading to a decreasing trust of the final users, such as clinicians. For this reason, when a model is deployed in practice, it would be important to advise users when the model's predictions may be unreliable, especially in high-stakes applications, including those in healthcare. Yet, reliability assessment of each machine learning prediction is still poorly addressed. Here, we review approaches that can support the identification of unreliable predictions, we harmonize the notation and terminology of relevant concepts, and we highlight and extend possible interrelationships and overlap among concepts. We then demonstrate, on simulated and real data for ICU in-hospital death prediction, a possible integrative framework for the identification of reliable and unreliable predictions. To do so, our proposed approach implements two complementary principles, namely the density principle and the local fit principle. The density principle verifies that the instance we want to evaluate is similar to the training set. The local fit principle verifies that the trained model performs well on training subsets that are more similar to the instance under evaluation. Our work can contribute to consolidating work in machine learning especially in medicine.