Database: The Journal of Biological Databases and Curation最新文献

Federated learning is an upcoming machine learning paradigm which allows data from multiple sources to be used for training of classifiers without the data leaving the source it originally resides. This can be highly valuable for use cases such as medical research, where gathering data at a central location can be quite complicated due to privacy and legal concerns of the data. In such cases, federated learning has the potential to vastly speed up the research cycle. Although federated and central learning have been compared from a theoretical perspective, an extensive experimental comparison of performances and learning behavior still lacks. We have performed a comprehensive experimental comparison between federated and centralized learning. We evaluated various classifiers on various datasets exploring influences of different sample distributions as well as different class distributions across the clients. The results show similar performances under a wide variety of settings between the federated and central learning strategies. Federated learning is able to deal with various imbalances in the data distributions. It is sensitive to batch effects between different datasets when they coincide with location, similar to central learning, but this setting might go unobserved more easily. Federated learning seems to be robust to various challenges such as skewed data distributions, high data dimensionality, multiclass problems, and complex models. Taken together, the insights from our comparison gives much promise for applying federated learning as an alternative to sharing data. Code for reproducing the results in this work can be found at: https://github.com/swiergarst/FLComparison.

Federated learning is an upcoming machine learning paradigm which allows data from multiple sources to be used for training of classifiers without the data leaving the source it originally resides. This can be highly valuable for use cases such as medical research, where gathering data at a central location can be quite complicated due to privacy and legal concerns of the data. In such cases, federated learning has the potential to vastly speed up the research cycle. Although federated and central learning have been compared from a theoretical perspective, an extensive experimental comparison of performances and learning behavior still lacks. We have performed a comprehensive experimental comparison between federated and centralized learning. We evaluated various classifiers on various datasets exploring influences of different sample distributions as well as different class distributions across the clients. The results show similar performances under a wide variety of settings between the federated and central learning strategies. Federated learning is able to deal with various imbalances in the data distributions. It is sensitive to batch effects between different datasets when they coincide with location, similar to central learning, but this setting might go unobserved more easily. Federated learning seems to be robust to various challenges such as skewed data distributions, high data dimensionality, multiclass problems, and complex models. Taken together, the insights from our comparison gives much promise for applying federated learning as an alternative to sharing data. Code for reproducing the results in this work can be found at: https://github.com/swiergarst/FLComparison.

Variant interpretation guidelines guide the process of determining the role of DNA variants in patients' health. Currently, hundreds of guidelines exist, each applicable to a particular clinical domain. However, they are scattered across multiple resources and scientific literature. To address this issue, we present VarGuideAtlas, a comprehensive repository of variant interpretation guidelines that compiles information from ClinGen, ClinVar, and PubMed. Our repository offers a user-friendly web interface with advanced search capabilities, enabling clinicians and researchers to efficiently find relevant guidelines tailored to specific genes, diseases, or variant types. We employ ontologies to characterize each guideline, ensuring consistency and improving interoperability with bioinformatics tools. VarGuideAtlas represents a significant advance toward standardizing variant interpretation practices, facilitating more informed decision-making, improved clinical outcomes, and more precise genomic research. VarGuideAtlas is publicly accessible via a web-based platform (https://genomics-hub.pros.dsic.upv.es:3016/).

Variant interpretation guidelines guide the process of determining the role of DNA variants in patients' health. Currently, hundreds of guidelines exist, each applicable to a particular clinical domain. However, they are scattered across multiple resources and scientific literature. To address this issue, we present VarGuideAtlas, a comprehensive repository of variant interpretation guidelines that compiles information from ClinGen, ClinVar, and PubMed. Our repository offers a user-friendly web interface with advanced search capabilities, enabling clinicians and researchers to efficiently find relevant guidelines tailored to specific genes, diseases, or variant types. We employ ontologies to characterize each guideline, ensuring consistency and improving interoperability with bioinformatics tools. VarGuideAtlas represents a significant advance toward standardizing variant interpretation practices, facilitating more informed decision-making, improved clinical outcomes, and more precise genomic research. VarGuideAtlas is publicly accessible via a web-based platform (https://genomics-hub.pros.dsic.upv.es:3016/).

Genome editing (GE) is widely recognized as an effective and valuable technology in life sciences research. However, certain genes are difficult to edit depending on some factors such as the type of species, sequences, and GE tools. Therefore, confirming the presence or absence of GE practices in previous publications is crucial for the effective designing and establishment of research using GE. Although the Genome Editing Meta-database (GEM: https://bonohu.hiroshima-u.ac.jp/gem/) aims to provide as comprehensive GE information as possible, it does not indicate how each registered gene is involved in GE. In this study, we developed a systematic method for extracting essential GE information using large language models from the information based on GEM and GE-related articles. This approach allows for a systematic and efficient investigation of GE information that cannot be achieved using the current GEM alone. In addition, by converting the extracted GE information into metrics, we propose a potential application of this method to prioritize genes for future research. The extracted GE information and novel GE-related scores are expected to facilitate the efficient selection of target genes for GE and support the design of research using GE. Database URLs:  https://github.com/szktkyk/extract_geinfo, https://github.com/szktkyk/visualize_geinfo.

Genome editing (GE) is widely recognized as an effective and valuable technology in life sciences research. However, certain genes are difficult to edit depending on some factors such as the type of species, sequences, and GE tools. Therefore, confirming the presence or absence of GE practices in previous publications is crucial for the effective designing and establishment of research using GE. Although the Genome Editing Meta-database (GEM: https://bonohu.hiroshima-u.ac.jp/gem/) aims to provide as comprehensive GE information as possible, it does not indicate how each registered gene is involved in GE. In this study, we developed a systematic method for extracting essential GE information using large language models from the information based on GEM and GE-related articles. This approach allows for a systematic and efficient investigation of GE information that cannot be achieved using the current GEM alone. In addition, by converting the extracted GE information into metrics, we propose a potential application of this method to prioritize genes for future research. The extracted GE information and novel GE-related scores are expected to facilitate the efficient selection of target genes for GE and support the design of research using GE. Database URLs:  https://github.com/szktkyk/extract_geinfo, https://github.com/szktkyk/visualize_geinfo.

Gymnosperms are a clade of non-flowering plants that include about 1000 living species. Due to their complex genomes and lack of genomic resources, functional annotation in genomics and transcriptomics on gymnosperms suffers from limitations. Here we present gymnotoa-db, which is a novel, publicly accessible relational database designed to facilitate functional annotation in gymnosperms. This database stores non-redundant records of gymnosperm proteins, encompassing taxonomic and functional information. The complementary software, gymnotoa-app, enables users to download gymnotoa-db and execute a comprehensive functional annotation pipeline for high-throughput sequencing-derived DNA or cDNA sequences. gymnotoa-app's user-friendly interface and efficient algorithms streamline the functional annotation process, making it an invaluable tool for researchers studying gymnosperms. We compared gymnotoa-app's performance against other annotation tools utilizing disparate reference databases. Our results demonstrate gymnotoa-app's superior ability to accurately annotate gymnosperm transcripts, recovering a greater number of transcripts and unique, non-redundant Gene Ontology terms. gymnotoa-db's distinctive features include comprehensive coverage with a non-redundant dataset of gymnosperm protein sequences, robust functional information that integrates data from multiple ontology systems, including GO, KEGG, EC, and MetaCYC, while keeping the taxonomic context, including Arabidopsis homologs. Database URL: https://blogs.upm.es/gymnotoa-db/2024/09/19/gymnotoa-app/.

Gymnosperms are a clade of non-flowering plants that include about 1000 living species. Due to their complex genomes and lack of genomic resources, functional annotation in genomics and transcriptomics on gymnosperms suffers from limitations. Here we present gymnotoa-db, which is a novel, publicly accessible relational database designed to facilitate functional annotation in gymnosperms. This database stores non-redundant records of gymnosperm proteins, encompassing taxonomic and functional information. The complementary software, gymnotoa-app, enables users to download gymnotoa-db and execute a comprehensive functional annotation pipeline for high-throughput sequencing-derived DNA or cDNA sequences. gymnotoa-app's user-friendly interface and efficient algorithms streamline the functional annotation process, making it an invaluable tool for researchers studying gymnosperms. We compared gymnotoa-app's performance against other annotation tools utilizing disparate reference databases. Our results demonstrate gymnotoa-app's superior ability to accurately annotate gymnosperm transcripts, recovering a greater number of transcripts and unique, non-redundant Gene Ontology terms. gymnotoa-db's distinctive features include comprehensive coverage with a non-redundant dataset of gymnosperm protein sequences, robust functional information that integrates data from multiple ontology systems, including GO, KEGG, EC, and MetaCYC, while keeping the taxonomic context, including Arabidopsis homologs. Database URL: https://blogs.upm.es/gymnotoa-db/2024/09/19/gymnotoa-app/.

Mitigating the effects of environmental exploitation on forests requires robust data analysis tools to inform sustainable management strategies and enhance ecosystem resilience. Access to extensive, integrated plant biodiversity data, spanning decades, is essential for this purpose. However, such data are often fragmented across diverse datasets with varying standards, posing two key challenges: first, integrating these datasets into a unified, well-structured data warehouse, and second, handling the vast volume of data using big data technologies to analyze and monitor the temporal evolution of ecosystems. To address these challenges, we developed and used an extract, transform, and load (ETL) protocol that curated and integrates 4482 forestry datasets from around the world, dating back to the 18th century, into a 100-GB data warehouse containing over 172 million records sourced from the Global Biodiversity Information Facility repository. We implemented Python scripts and a NoSQL MongoDB database to streamline and automate the ETL process, using the data warehouse to create the ForestForward web platform. ForestForward is a free, user-friendly application developed using the Django framework, which enables users to consult, download, and visualize the curated data. The platform allows users to explore data layers by year and observe the temporal evolution of ecosystems through visual representations. Database URL: https://forestforward.udl.cat.

Mitigating the effects of environmental exploitation on forests requires robust data analysis tools to inform sustainable management strategies and enhance ecosystem resilience. Access to extensive, integrated plant biodiversity data, spanning decades, is essential for this purpose. However, such data are often fragmented across diverse datasets with varying standards, posing two key challenges: first, integrating these datasets into a unified, well-structured data warehouse, and second, handling the vast volume of data using big data technologies to analyze and monitor the temporal evolution of ecosystems. To address these challenges, we developed and used an extract, transform, and load (ETL) protocol that curated and integrates 4482 forestry datasets from around the world, dating back to the 18th century, into a 100-GB data warehouse containing over 172 million records sourced from the Global Biodiversity Information Facility repository. We implemented Python scripts and a NoSQL MongoDB database to streamline and automate the ETL process, using the data warehouse to create the ForestForward web platform. ForestForward is a free, user-friendly application developed using the Django framework, which enables users to consult, download, and visualize the curated data. The platform allows users to explore data layers by year and observe the temporal evolution of ecosystems through visual representations. Database URL: https://forestforward.udl.cat.