Biology subjects

Medina-Ortiz, D.

Publications and source records attributed to Medina-Ortiz, D..

10 recordsLinked to original sources

bioRxiv2026

PhageLysData: an evidence-aware and AI-ready dataset of phage lytic enzymes and depolymerases

Bacteriophage lytic enzymes and depolymerases are relevant to phage biology, antimicrobial development, and protein engineering, but their sequence and annotation data remain dispersed across general databases, specialized resources, genome-centred collections, and prediction-oriented datasets. We present PhageLysData, an evidence-aware and AI-ready resource constructed through reproducible multisource integration, provenance tracking, and exact-sequence consolidation. The release integrates 807,366 source observations from seven primary resources into 759,105 unique exact-sequence entities, comprising an evidence-supported Core of 11,867 entities, a Prediction Extension of 745,092 prediction-only candidates, and 2,146 Context entities retained for provenance and reference. This architecture preserves broad sequence-space coverage while maintaining a clear distinction between non-predictive and prediction-derived support. Core entities are enriched with harmonized biological annotations, physicochemical properties, independent InterProScan-derived functional annotations, mapped PDB and AlphaFold DB structural assets, and reusable numerical representations. For 11,259 eligible Core sequences, PhageLysData provides embeddings from 11 protein language models together with one-hot encoding under a common representation contract. Release-facing examples demonstrate latent-space exploration, unsupervised clustering, supervised classification, and evidence-aware candidate retrieval without defining a universal predictive benchmark. PhageLysData provides a traceable, versioned, and computationally accessible foundation for protein retrieval, comparative analysis, task-specific dataset construction, and machine-learning applications involving phage lytic enzymes and depolymerases.

bioinformatics↗

bioRxiv2026

SilkRoute: A Descriptor-Driven Framework for Reproducible Multi-Source Biomolecular Data Acquisition

BackgroundBiomolecular dataset construction often requires coordinated retrieval from heterogeneous repositories, identifier mapping, cross-reference enrichment, source-specific parsing, and provenance recording. These operations are frequently implemented through project-specific scripts, making acquisition procedures difficult to inspect, reproduce, or adapt across studies. We present SilkRoute, an open-source Python framework that formalizes biomolecular data acquisition as descriptor-defined, source-aware, and provenance-tracked workflows, providing a reproducible foundation for multi-source biomolecular dataset construction. ResultsSilkRoute uses machine-readable YAML descriptors to specify dataset intent, biomolecular modality, workflow mode, query logic, enrichment resources, execution parameters, and export settings. These descriptors drive a common execution model that coordinates primary retrieval and downstream enrichment while preserving source-specific outputs, interaction evidence when available, the original workflow configuration, metadata, and run summaries. We evaluated this model through three representative acquisition scenarios spanning proteins, compounds, and molecular interactions. In the protein-centered workflow, SilkRoute retrieved 2,444 reviewed antimicrobial protein records from UniProt and generated complementary outputs from AlphaFold DB, InterPro, Pathway Commons, and the Protein Data Bank. In the compound-centered workflow, a ChEMBL IC50 query produced 1,445,939 activity records organized into query-defined potency ranges. In the interaction-centered workflow, 2,253 UniProt protein records were expanded with 902,713 BioGRID interaction records and 5,702 STRING interaction-partner records. Across these scenarios, the framework successfully applied the same descriptor-defined acquisition model to distinct biomolecular entity types, retrieval strategies, enrichment paths, and output structures. ConclusionsSilkRoute extends beyond sequence retrieval by providing a reusable acquisition layer for constructing multi-source biomolecular datasets. By separating primary retrieval from enrichment and preserving source-aware outputs together with workflow descriptors and execution metadata, the framework makes acquisition procedures easier to inspect, reproduce, archive, and adapt. SilkRoute does not replace biological curation, label validation, deduplication, partitioning, or benchmarking, but provides structured and traceable acquisition packages that support these downstream processes.

bioinformatics↗

bioRxiv2026

Data-Centric Evaluation of Protein Function Prediction Pipelines

Performance estimates in protein function prediction depend not only on model choice but also on upstream decisions that define the learning problem. Using antioxidant protein classification as a controlled case study, we evaluated how dataset harmonisation, protein representation, redundancy control, and partitioning strategy affect protein machine learning pipelines. We integrated 18,804 records from 12 publicly available dataset entries into a curated consensus dataset of 4,193 protein sequences. One-hot encoding and six pretrained protein language model representations were evaluated as model inputs and as similarity spaces for redundancy reduction and distance-aware splitting. Representation choice substantially altered dataset geometry, retained dataset size, class balance, and downstream evaluation. At representation-specific p90 thresholds, one-hot encoding retained the complete dataset, whereas pretrained embeddings retained between 5% and 25% of sequences. Distance-aware partitioning reduced apparent performance relative to random splitting by up to 0.15 MCC before redundancy control, while this difference narrowed after similarity filtering. Selected configurations nevertheless maintained high performance under stricter evaluation, reaching an MCC of 0.84. These findings show that performance estimates should be interpreted as outcomes of complete data-centric workflows rather than isolated properties of predictive models.

bioinformatics↗

bioRxiv2026

MAOMAO: An Ontology-Guided FAIR Resource for Harmonized Peptide Toxicity Data

Peptide toxicity is a critical safety and developability parameter in peptide discovery and therapeutic development, yet relevant information remains fragmented across databases, literature resources, and curated datasets. Here, we present MAOMAO, an ontology-guided FAIR-oriented resource that integrates and harmonizes peptide toxicity data from 54 sources. MAOMAO contains 71,701 unique peptide sequences across seven toxicity-related endpoints, represented as 501,907 sequence endpoint combinations with endpoint-specific evidence states and explicit encoding of unavailable information. The resource combines standardized terminology, a hierarchical toxicity vocabulary, evidence-aware state resolution, provenance-aware metadata, 41 physicochemical descriptors, 10 protein language model representations, and a one-hot baseline. It provides endpoint-specific benchmark partitions across splitting strategies and random seeds, reusable with numerical representations. MAOMAO establishes a reusable framework for peptide toxicity research and data-driven toxicology.

pharmacology and toxicology↗

bioRxiv2026

Integrated heart rate variability and physiological profiling reveals autonomic phenotypes in older adults from a high-southern-latitude population

Heart rate variability (HRV) is widely used to assess autonomic regulation, but its interpretation in older adults is influenced by age, sex, body composition, and hemodynamic status, particularly in underrepresented populations living in geographically extreme environments. We analyzed 530 community-dwelling older adults from the Magallanes region in southern Chile using an integrated framework that combined HRV indices with demographic, anthropometric, and cardiovascular descriptors. After quality-controlled preprocessing, we characterized the distribution and association structure of autonomic and physiological variables and then performed a large-scale unsupervised clustering benchmark across multiple feature spaces, dimensionality-reduction strategies, and clustering algorithms. Conventional descriptors explained only a limited proportion of HRV variability, whereas integrated multivariate analysis revealed a structured continuum of autonomic heterogeneity. A six-cluster solution provided the best compromise between separation, balance, and physiological interpretability, identifying profiles that differed in HRV magnitude, blood pressure burden, body composition, sex distribution, and age structure. These findings indicate that autonomic regulation in older adults cannot be adequately summarized by isolated descriptors such as age, body mass index, or blood pressure alone. Instead, it is better represented as a multidimensional physiological organization that supports future hypothesis generation for risk stratification and longitudinal monitoring in aging populations.

neuroscience↗

bioRxiv2026

Visualize, Explore, and Select: A protein Language Model-based Approach Enabling Navigation of Protein Sequence Space for Enzyme Discovery and Mining

The rapid expansion of protein sequence databases continues to outpace functional characterization, creating a persistent bottleneck in enzyme discovery and mining--particularly in large, heterogeneous, and sparsely annotated sequence spaces. This gap is amplified by visualization challenges and the lack of informed strategies for exploration, selection, and mining across sequence spaces. Here, we present an embedding-based workflow, implemented in a computational platform called SelectZyme, for alignment-free visualization and exploration of protein sequence space that combines protein language model (pLM) representations with dimensionality reduction and hierarchical density-based clustering. The approach links complementary visualizations of protein sequence space as low-dimensional landscapes, connectivity projections (minimum spanning trees), and dendrogram-based organization, enabling coherent interactive exploration and candidate selection across global context and local neighborhoods without relying on sequence-identity thresholds, EC numbers, conserved motifs, or predefined functional annotations. Across distinct case studies, we demonstrate that embedding-defined neighborhoods remain structurally conserved even when sequence identity falls within the twilight zone, and that coherent functional organization emerges also for modular protein segments in a fully unsupervised analysis. We also show how this workflow supports user friendly, interactive and scalable enzyme mining in a sparsely annotated, complex multi-family protein space surpassing > 100,000 sequences, enabling constrained candidate selection around experimentally validated anchors. By enabling interactive exploration across visualizations, supporting informed candidate selection, our workflow streamlines biocatalyst discovery and helps to bridge uncharacterized sequence-space to functional characterization campaigns - thus providing a broad starting point to downstream protein engineering, machine-learning-guided design cycles, and iterative experimental screening campaigns.

bioinformatics↗

bioRxiv2025

Protein language models accelerate the discovery of Plastic-Degrading Enzymes

Plastic pollution presents a critical environmental challenge, necessitating innovative and sustainable solutions. In this context, biodegradation using microorganisms and enzymes offers an environmentally friendly alternative. This work introduces an AI-driven frame-work that integrates machine learning (ML) and generative models to accelerate the discovery and design of plastic-degrading enzymes. By leveraging pre-trained protein language models and curated datasets, we developed seven ML-based binary classification models to identify enzymes targeting specific plastic substrates, achieving an average accuracy of 89%. The framework was applied to over 6,000 enzyme sequences from the RemeDB to classify enzymes targeting diverse plastics, including PET, PLA, and Nylon. Besides, generative learning strategies combined with trained classification models in this work were applied for de novo generation of PET-degrading enzymes. Structural bioinformatics validated potential candidates through in-silico analysis, highlighting differences in physicochemical properties between generated and experimentally validated enzymes. Moreover, generated sequences exhibited lower molecular weights and higher aliphatic indices, features that may enhance interactions with hydrophobic plastic substrates. These findings highlight the utility of AI-based approaches in enzyme discovery, providing a scalable and efficient tool for addressing plastic pollution. Future work will focus on experimental validation of promising candidates and further refinement of generative strategies to optimize enzymatic performance.

bioengineering↗

bioRxiv2024

Peptipedia v2.0: A peptide sequence database and user-friendly web platform. A major update

In recent years, peptides have gained significant relevance due to their therapeutic properties. The surge in peptide production and synthesis has generated vast amounts of data, enabling the creation of comprehensive databases and information repositories. Advances in sequencing techniques and artificial intelligence have further accelerated the design of tailor-made peptides. However, leveraging these techniques requires versatile and continuously updated storage systems, along with tools that facilitate peptide research and the implementation of machine learning for predictive systems. This work introduces Peptipedia v2.0, one of the most comprehensive public repositories of peptides, supporting biotechnological research by simplifying peptide study and annotation. Peptipedia v2.0 has expanded its collection by over 45% with peptide sequences that have reported biological activities. The functional biological activity tree has been revised and enhanced, incorporating new categories such as cosmetic and dermatological activities, molecular binding, and anti-ageing properties. Utilizing protein language models and machine learning, more than 90 binary classification models have been trained, validated, and incorporated into Peptipedia v2.0. These models exhibit average sensitivities and specificities of 0.877 {+/-} 0.0530 and 0.873 {+/-}0.054, respectively, facilitating the annotation of more than 3.6 million peptide sequences with unknown biological activities, also registered in Peptipedia v2.0. Additionally, Peptipedia v2.0 introduces description tools based on structural and ontological properties and user-friendly machinelearning tools to facilitate the application of machine-learning strategies to study peptide sequences. Peptipedia v2.0 is accessible under the Creative Commons CC BY-NC-ND 4.0 license at https://peptipedia.cl/.

bioinformatics↗

bioRxiv2024

Interpretable and explainable predictive machine learning models for data-driven protein engineering

Protein engineering using directed evolution and (semi)rational design has emerged as a powerful strategy for optimizing and enhancing enzymes or proteins with desired properties. Integrating artificial intelligence methods has further enhanced and accelerated protein engineering through predictive models developed in data-driven strategies. However, the lack of explainability and interpretability in these models poses challenges. Explainable Artificial Intelligence addresses the interpretability and explainability of machine learning models, providing transparency and insights into predictive processes. Nonetheless, there is a growing need to incorporate explainable techniques in predicting protein properties in machine learning-assisted protein engineering. This work explores incorporating explainable artificial intelligence in predicting protein properties, emphasizing its role in trustworthiness and interpretability. It assesses different machine learning approaches, introduces diverse explainable methodologies, and proposes strategies for seamless integration, improving trust-worthiness. Practical cases demonstrate the explainable models effectiveness in identifying DNA binding proteins and optimizing Green Fluorescent Protein brightness. The study highlights the utility of explainable artificial intelligence in advancing computationally assisted protein design, fostering confidence in model reliability.

bioinformatics↗

bioRxiv2024

RUDEUS, a machine learning classification system to study DNA-Binding proteins

DNA-binding proteins are essential in different biological processes, including DNA replication, transcription, packaging, and chromatin remodelling. Exploring their characteristics and functions has become relevant in diverse scientific domains. Computational biology and bioinformatics have assisted in studying DNA-binding proteins, complementing traditional molecular biology methods. While recent advances in machine learning have enabled the integration of predictive systems with bioinformatic approaches, there still needs to be generalizable pipelines for identifying unknown proteins as DNA-binding and assessing the specific type of DNA strand they recognize. In this work, we introduce RUDEUS, a Python library featuring hierarchical classification models designed to identify DNA-binding proteins and assess the specific interaction type, whether single-stranded or double-stranded. RUDEUS has a versatile pipeline capable of training predictive models, synergizing protein language models with supervised learning algorithms, and integrating Bayesian optimization strategies. The trained models have high performance, achieving a precision rate of 95% for DNA-binding identification and 89% for discerning between single-stranded and doublestranded interactions. RUDEUS includes an exploration tool for evaluating unknown protein sequences, annotating them as DNA-binding, and determining the type of DNA strand they recognize. Moreover, a structural bioinformatic pipeline has been integrated into RUDEUS for validating the identified DNA strand through DNA-protein molecular docking. These comprehensive strategies and straightforward implementation demonstrate comparable performance to high-end models and enhance usability for integration into protein engineering pipelines.

bioinformatics↗