A5047489800- Biomedical Text Mining and Ontologies
- Topic Modeling
- Natural Language Processing Techniques
- Semantic Web and Ontologies
- Cell Image Analysis Techniques
- Bioinformatics and Genomic Networks
- Advanced Proteomics Techniques and Applications
- Metabolomics and Mass Spectrometry Studies
- Scientific Computing and Data Management
- Advanced Biosensing Techniques and Applications
- scientometrics and bibliometrics research
- Epigenetics and DNA Methylation
- Advanced Text Analysis Techniques
- Research Data Management Practices
- SARS-CoV-2 and COVID-19 Research
- Advanced Graph Neural Networks
- Biotin and Related Studies
- Cancer Immunotherapy and Biomarkers
- Cellular Mechanics and Interactions
- Recommender Systems and Techniques
- Digital Imaging for Blood Diseases
- Single-cell and spatial transcriptomics
- Long-Term Effects of COVID-19
- Topological and Geometric Data Analysis
- COVID-19 Clinical Research Studies
European Bioinformatics Institute
2019-2023
Wellcome Trust
2019-2023
Heidelberg Institute for Theoretical Studies
2019-2021
Universidad del Valle
2012-2018
Asociación por los Derechos Civiles
2016
Improved understanding and management of COVID-19, a potentially life-threatening disease, could greatly reduce the threat posed by its etiologic agent, SARS-CoV-2. Toward this end, we have identified core peripheral blood immune signature across 63 hospital-treated patients with COVID-19 who were otherwise highly heterogeneous. The includes discrete changes in B myelomonocytic cell composition, profoundly altered T phenotypes, selective cytokine/chemokine upregulation SARS-CoV-2-specific...
Abstract The International Mouse Phenotyping Consortium (IMPC; https://www.mousephenotype.org/) web portal makes available curated, integrated and analysed knockout mouse phenotyping data generated by the IMPC project consisting of 85M points over 95,000 statistically significant phenotype hits mapped to human diseases. delivers a substantial reference dataset that supports enrichment various domain-specific projects databases, as well wider research clinical community, where...
Scientific research relies on computer software, yet software is not always developed following practices that ensure its quality and sustainability. This manuscript does aim to propose new development best practices, but rather provide simple recommendations encourage the adoption of existing practices. Software promote better improves reproducibility reusability research. These are designed around Open Source values, practical suggestions contribute making source code more discoverable,...
This article presents new alternatives to the similarity function for TextRank algorithm automatic summarization of texts. We describe generalities and different functions we propose. Some these variants achieve a significative improvement using same metrics dataset as original publication.
A Correction to this paper has been published: https://doi.org/10.1038/s41591-020-01186-5
Abstract Person-to-person transmission of SARS-CoV-2 virus has triggered a global emergency because its potential to cause life-threatening Covid-19 disease. By comparison paucisymptomatic clearance by most individuals, been proposed reflect insufficient and/or pathologically exaggerated immune responses. Here we identify consensus peripheral blood signature across 63 hospital-treated patients who were otherwise highly heterogeneous. The core conspicuously blended adaptive B cell responses...
How can we represent hierarchical information present in large type inventories for entity typing? We study the suitability of hyperbolic embeddings to capture relations between mentions context and their target types a shared vector space. evaluate on two datasets propose different techniques extract from inventory: an expert-generated ontology by automatically mining dataset. The model shows improvements some but not all cases over its Euclidean counterpart. Our analysis suggests that...
An experimental protocol is a sequence of tasks and operations executed to perform research in biological biomedical areas, e.g. biology, genetics, immunology, neurosciences, virology. Protocols often include references equipment, reagents, descriptions critical steps, troubleshooting tips, as well any other information that researchers deem important for facilitating the reusability protocol. Although protocols are central reproducibility, cursory. There need unified framework with respect...
Abstract PDCM Finder (www.cancermodels.org) is a cancer research platform that aggregates clinical, genomic and functional data from patient-derived xenografts, organoids cell lines. It was launched in April 2022 as successor of the PDX portal, which focused solely on xenograft models. Currently portal has over 6200 models across 13 types, including rare paediatric (17%) minority ethnic backgrounds (33%), making it largest free to consumer open access resource this kind. The standardises,...
An amendment to this paper has been published and can be accessed via a link at the top of paper.
Abstract Motivation High-throughput phenomic projects generate complex data from small treatment and large control groups that increase the power of analyses but introduce variation over time. A method is needed to utlize a set temporally local controls maximizes analytic while minimizing noise unspecified environmental factors. Results Here we ‘soft windowing’, methodological approach selects window time includes most appropriate for analysis. Using phenotype International Mouse Phenotyping...
A significant portion of biomedical literature is represented in a manner that makes it difficult for consumers to find or aggregate content through computational query. One approach facilitate reuse the scientific structure this information as linked data using standardized web technologies. In paper we present second version Biotea, semantic, open-access subset PubMed Central has been enhanced with specialized annotation pipelines uses existing infrastructure from National Center...
Medical images contains valuable information that is not explicit and readable for the machine. For instance, an image may contain about anatomy abnormal structures. However, this kind of can only be interpreted by a medical domain expert. This paper proposes SMITag, collaborative semantic annotation tool combines features DICOM Viewer together with social network, so consensus experts makes easier enrichment tasks, sorting retrieval.
A significant portion of biomedical literature is represented in a manner that makes it difficult for consumers to find or aggregate content through computational query. One approach facilitate reuse the scientific structure this information as linked data using standardized web technologies. In paper we present second version Biotea, semantic, open-access subset PubMed Central has been enhanced with specialized annotation pipelines uses existing infrastructure from National Center...
This paper introduces a simple and effective form of data augmentation for recommender systems. A paraphrase similarity model is applied to widely available textual data, such as reviews product descriptions, yielding new semantic relations that are added the user-item graph. increases density graph without needing further labeled data. The evaluated on variety recommendation algorithms, using Euclidean, hyperbolic, complex spaces, over three categories Amazon with differing characteristics....
Learning faithful graph representations as sets of vertex embeddings has become a fundamental intermediary step in wide range machine learning applications. We propose the systematic use symmetric spaces representation learning, class encompassing many previously used embedding targets. This enables us to introduce new method, Finsler metrics integrated Riemannian optimization scheme, that better adapts dissimilar structures graph. develop tool analyze and infer structural properties data...