1,354,163 research outputs found
Analysis and Visualization of Mutation Enrichments for Selected Genomic Regions and Cancer Types
No full textSeveral studies highlight the relevance of somatic mutations in non-coding regions of the genome which exhibit common interesting behaviors. MutViz is a tool for the identification of mutation enrichments on arbitrary sets of user-defined regions; for a variety of cancer types, it contains preloaded mutations from public datasets, well organized within an effective database organization. MutViz provides a user-friendly interface helping the user in providing sets of regions as input and in obtaining their fast exploration as output, together with simple statistical testing of novel hypotheses
Exploring genomic datasets: From batch to interactive and back
No full textGenomic data management is focused on achieving high performance over big datasets using batch, cloud-based architectures; this enables the execution of massive pipelines, but hampers the capability of exploring the solution space when it is not well-defined, by choosing different experimental samples or query extraction parameters. We present PyGMQL, a Python-based interoperability software layer that enables testing of experimental pipelines; PyGMQL solves the impedance mismatch between a batch execution environment and the agile programming style of Python, and provides transparency of access when exploration requires integrating local and remote resources.Wrapping PyGMQLand Python primitives within Jupyter notebooks guarantees reproducibility of the pipeline when used in different contexts or by different scientists. The software is freely available at https://github.com/DEIB-GECO/PyGMQL
Ontology-Driven Metadata Enrichment for Genomic Datasets
No full textData-driven genomic research requires accessing several repositories of genomic datasets, produced by international consortia, which provide open access to extremely valuable and well curated biological content. The associated metadata, describing experimental and biological conditions, are highly heterogeneous; consequently, dataset collection and integration is difficult – it requires data conversions and term matching which needs to be done by humans, with biological expertise. In this paper, we present a method and tools for ontology-driven metadata enrichment. We select few relevant features which are provided by most repositories, and then we comparatively evaluate several search services providing ontological access, eventually associating each feature with the specific ontologies which are most suited to describe them. We also provide an expert validation of the approach. The method and tools are deployed in a large repository of open data, which will be soon available to the research community
Processing genome-wide association studies within a repository of heterogeneous genomic datasets
No full textBackground
Genome Wide Association Studies (GWAS) are based on the observation of genome-wide sets of genetic variants – typically single-nucleotide polymorphisms (SNPs) – in different individuals that are associated with phenotypic traits. Research efforts have so far been directed to improving GWAS techniques rather than on making the results of GWAS interoperable with other genomic signals; this is currently hindered by the use of heterogeneous formats and uncoordinated experiment descriptions.
Results
To practically facilitate integrative use, we propose to include GWAS datasets within the META-BASE repository, exploiting an integration pipeline previously studied for other genomic datasets that includes several heterogeneous data types in the same format, queryable from the same systems. We represent GWAS SNPs and metadata by means of the Genomic Data Model and include metadata within a relational representation by extending the Genomic Conceptual Model with a dedicated view. To further reduce the gap with the descriptions of other signals in the repository of genomic datasets, we perform a semantic annotation of phenotypic traits. Our pipeline is demonstrated using two important data sources, initially organized according to different data models: the NHGRI-EBI GWAS Catalog and FinnGen (University of Helsinki). The integration effort finally allows us to use these datasets within multisample processing queries that respond to important biological questions. These are then made usable for multi-omic studies together with, e.g., somatic and reference mutation data, genomic annotations, epigenetic signals.
Conclusions
As a result of our work on GWAS datasets, we enable 1) their interoperable use with several other homogenized and processed genomic datasets in the context of the META-BASE repository; 2) their big data processing by means of the GenoMetric Query Language and associated system. Future large-scale tertiary data analysis may extensively benefit from the addition of GWAS results to inform several different downstream analysis workflows
Designing and Evaluating Deep Learning Models for Cancer Detection on Gene Expression Data
No full textTranscription profiling enables researchers to understand the activity of the genes in various experimental conditions; in human genomics, abnormal gene expression is typically correlated with clinical conditions. An important application is the detection of genes which are most involved in the development of tumors, by contrasting normal and tumor cells of the same patient. Several statistical and machine learning techniques have been applied to cancer detection; more recently, deep learning methods have been attempted, but they have typically failed in meeting the same performance as classical algorithms. In this paper, we design a set of deep learning methods that can achieve similar performance as the best machine learning methods thanks to the use of external information or of data augmentation; we demonstrate this result by comparing the performance of new methods against several baselines
Integration of biomolecular interaction data in a genomic and proteomic data warehouse to support biomedical knowledge discovery
No full textThe growing available genomic and proteomic information gives new opportunities for novel research approaches and biomedical discoveries through effective data management and analysis support. Integration and comprehensive evaluation of available controlled data can highlight information patterns leading to unveil new biomedical knowledge. For this purpose, the University Politecnico di Milano, is developing a software framework to create and maintain a Genomic and Proteomic Data Warehouse (GPDW) that integrates information from many data sources on the basis of a conceptual data model that relates molecular entities and biomedical features. Here we illustrate and discuss the extension of framework for integrating biomolecular interaction data in the GPDW. The comprehensive and mining of the reliable interaction data together with the other biomolecular information in the GPDW constitutes a powerful computational support for novel biomedical knowledge discoveries
PyGMQL: scalable data extraction and analysis for heterogeneous genomic datasets
No full textBackground: With the growth of available sequenced datasets, analysis of heterogeneous processed data can answer increasingly relevant biological and clinical questions. Scientists are challenged in performing efficient and reproducible data extraction and analysis pipelines over heterogeneously processed datasets. Available software packages are suitable for analyzing experimental files from such datasets one by one, but do not scale to thousands of experiments. Moreover, they lack proper support for metadata manipulation. Results: We present PyGMQL, a novel software for the manipulation of region-based genomic files and their relative metadata, built on top of the GMQL genomic big data management system. PyGMQL provides a set of expressive functions for the manipulation of region data and their metadata that can scale to arbitrary clusters and implicitly apply to thousands of files, producing millions of regions. PyGMQL provides data interoperability, distribution transparency and query outsourcing. The PyGMQL package integrates scalable data extraction over the Apache Spark engine underlying the GMQL implementation with native Python support for interactive data analysis and visualization. It supports data interoperability, solving the impedance mismatch between executing set-oriented queries and programming in Python. PyGMQL provides distribution transparency (the ability to address a remote dataset) and query outsourcing (the ability to assign processing to a remote service) in an orthogonal way. Outsourced processing can address cloud-based installations of the GMQL engine. Conclusions: PyGMQL is an effective and innovative tool for supporting tertiary data extraction and analysis pipelines. We demonstrate the expressiveness and performance of PyGMQL through a sequence of biological data analysis scenarios of increasing complexity, which highlight reproducibility, expressive power and scalability
Exploiting ladder networks for gene expression classification
No full textThe application of deep learning to biology is of increasing relevance, but it is difficult; one of the main difficulties is the lack of massive amounts of training data. However, some recent applications of deep learning to the classification of labeled cancer datasets have been successful. Along this direction, in this paper, we apply Ladder networks, a recent and interesting network model, to the binary cancer classification problem; our results improve over the state of the art in deep learning and over the conventional state of the art in machine learning; achieving such results required a careful adaptation of the available datasets and tuning of the network
Optimal design of microgrid-based resilient hybrid electric vehicle station by considering uncertainties
No full textDesigning a resilient hybrid electric vehicle station that integrates battery electric vehicle (BEV) charging and hydrogen refueling, supported by renewable energy sources and hybrid storage systems, is a forward-thinking approach to sustainable transportation infrastructure. This design not only improves energy security but also contributes to environmental sustainability. The system, in terms of local energy sources, consists of solar, wind, battery, and hydrogen storage systems. To design the system optimally, considering uncertainties arising from renewable generation variability, vehicle demand fluctuations, and electricity market prices, a two-stage stochastic programming model is utilized, integrating Conditional Value-at-Risk (CVaR) to explicitly address extreme tail risks linked to demand jumps and generation inconsistencies. We conduct a performance evaluation of the designed system under varying grid connection capacities, resilience constraints, and charger configurations. The analysis reveals significant trade-offs among system resilience, investment costs, and operational profitability.</p
- …
