Large-scale Repositories Research Articles

Clustering protein functional families at large scale with hierarchical approaches.

Proteins, fundamental to cellular activities, reveal their function and evolution through their structure and sequence. CATH functional families (FunFams) are coherent clusters of protein domain sequences in which the function is conserved across their members. The increasing volume and complexity of protein data enabled by large-scale repositories like MGnify or AlphaFold Database requires more powerful approaches that can scale to the size of these new resources. In this work, we introduce MARC and FRAN, two algorithms developed to build upon and address limitations of GeMMA/FunFHMMER, our original methods developed to classify proteins with related functions using a hierarchical approach. We also present CATH-eMMA, which uses embeddings or Foldseek distances to form relationship trees from distance matrices, reducing computational demands and handling various data types effectively. CATH-eMMA offers a highly robust and much faster tool for clustering protein functions on a large scale, providing a new tool for future studies in protein function and evolution.

Genomic analysis of hypoxia and mitophagy related genes with prognosis and characterization of the immune microenvironment in LUAD.

Background: Lung adenocarcinoma (LUAD) stands as a prominent subtype within the realm of non-small cell lung cancer and constitutes a primary contributor to cancer-related mortality on a global scale. Notably, hypoxia, a prevalent attribute within solid tumor environments, and mitophagy, a selective manifestation of autophagy dedicated to the removal of damaged mitochondria, have risen to prominence as pivotal factors influencing the initiation and advancement of tumorigenesis. Methods: This investigation harnessed publicly accessible genomic datasets encompassing LUAD patients to delineate genes linked to hypoxia and mitophagy, termed hereafter as hypoxia and mitophagy-related genes (HMRGs). Large-scale repositories furnished both gene expression profiles and clinical particulars. The expression profiles of HMRGs were meticulously scrutinized across 1,093 LUAD specimens, leveraging resources such as The Cancer Genome Atlas and Gene Expression Omnibus datasets. A methodical exploration of HMRG patterns within LUAD led to the discernment of two distinct molecular subtypes. Moreover, a discernible correlation emerged between the subtypes and their respective clinical attributes. A risk scoring system was formulated to prognosticate overall survival (OS) and therapeutic responsiveness in LUAD patients. Subsequently, the reliability of this scoring system was authenticated, and a nomogram was adopted to refine the clinical utility range of the risk score. The proliferation and migration impacts of KRT8 on LUAD cells were evaluated through cck8 assays, edu assays, and transwell assays, the results were further validated in vivo. Results: Elevated risk scores were indicative of unfavorable OS probabilities. Furthermore, these risk scores exhibited associations with immune checkpoints and chemotherapeutic drug sensitivity. Collectively, our exhaustive analysis of HMRGs in LUAD patients unveiled their conceivable participation in configuring the multifaceted tumor microenvironment, encompassing clinicopathological attributes and prognosis. A sequence of experiments illuminated the pro-proliferative and pro-migratory attributes of KRT8 in vitro and vivo, thus underscoring its carcinogenic potential. Conclusions: In this study, we have unearthed innovative gene signatures tethered to HMRGs, which harbor prognostic implications concerning patient outcomes. These insights hold potential for steering the development of targeted therapeutic modalities tailored for LUAD.

DNA Damage Atlas: an atlas of DNA damage and repair.

DNA damage and its improper repair are the major source of genomic alterations responsible for many human diseases, particularly cancer. To aid researchers in understanding the underlying mechanisms of genome instability, a number of genome-wide profiling approaches have been developed to monitor DNA damage and repair events. The rapid accumulation of published datasets underscores the critical necessity of a comprehensive database to curate sequencing data on DNA damage and repair intermediates. Here, we present DNA Damage Atlas (DDA, http://www.bioinformaticspa.com/DDA/), the first large-scale repository of DNA damage and repair information. Currently, DDA comprises 6,030 samples from 262 datasets by 59 technologies, covering 16 species, 10 types of damage and 135 treatments. Data collected in DDA was processed through a standardized workflow, including quality checks, hotspots identification and a series of feature characterization for the hotspots. Notably, DDA encompasses analyses of highly repetitive regions, ribosomal DNA and telomere. DDA offers a user-friendly interface that facilitates browsing, searching, genome browser visualization, hotspots comparison and data downloading, enabling convenient and thorough exploration for datasets of interest. In summary, DDA will stand as a valuable resource for research in genome instability and its association with diseases.

A large-scale repository of spoken narratives in French, German and Spanish from Cantonese-speaking learners

Interdisciplinary research on foreign language learning has important implications for learning and education. In this paper, we present the Repository of Third Language (L3) Spoken Narratives from Modern Language Learners in Hong Kong (L3HK Repository). This database contains 906 audio recordings and annotated transcripts of spoken narratives in French, German, and Spanish that were elicited from Cantonese-speaking (L1) young adults using a wordless picture book, “Frog, Where Are You?”. All participants spoke English as the second language (L2) and learned the target language as a third language (L3). We collected their demographic information, answers to a motivation questionnaire, parental socioeconomic status, and music background. Furthermore, for a subset of participants, we collected their L1 and L2 proficiency scores and additional experimental data on working memory and music perception. This database is valuable for examining cross-sectional changes in foreign language learning. The extensive data on phenotypes provide opportunities to explore learner-internal and learner-external factors in foreign language learning outcomes. These data may also be helpful for those who work on speech recognition.

DeGraphCS : Embedding Variable-based Flow Graph for Neural Code Search

With the rapid increase of public code repositories, developers maintain a great desire to retrieve precise code snippets by using natural language. Despite existing deep learning-based approaches that provide end-to-end solutions (i.e., accept natural language as queries and show related code fragments), the performance of code search in the large-scale repositories is still low in accuracy because of the code representation (e.g., AST) and modeling (e.g., directly fusing features in the attention stage). In this paper, we propose a novel learnable de ep G raph for C ode S earch (called deGraphCS ) to transfer source code into variable-based flow graphs based on an intermediate representation technique, which can model code semantics more precisely than directly processing the code as text or using the syntax tree representation. Furthermore, we propose a graph optimization mechanism to refine the code representation and apply an improved gated graph neural network to model variable-based flow graphs. To evaluate the effectiveness of deGraphCS , we collect a large-scale dataset from GitHub containing 41,152 code snippets written in the C language and reproduce several typical deep code search methods for comparison. The experimental results show that deGraphCS can achieve state-of-the-art performance and accurately retrieve code snippets satisfying the needs of the users.

The opportunities of research parasitism: A case study using the Barcode of Life Data System (BOLD).

The Barcode of Life Data System (BOLD) is primarily used to identify biological specimens based on a mitochondrial gene sequence and has been an underpinning resource for life science researchers. Importantly, curators of BOLD archive DNA extracts where possible, and also record contaminant sequences that can be made available on request. This collegial offering of samples and data led to our work describing the serendipitous discovery of new interactions between a Torix Rickettsia bacterium and their arthropod hosts and resulted in winning the 2022 Junior Research Parasite Award. A case study of this work is presented, which discusses the opportunities provided by secondary data and how careful maintenance of such large-scale repositories plays a vital role in scientific research that goes beyond obvious lines of enquiry.

The Chinese Lexicon Project II: A megastudy of speeded naming performance for 25,000+ traditional Chinese two-character words.

Using a megastudy approach, (Tse et al.,2017Behavior Research Methods, 49,1503-1519)established a large-scale repository of lexical variables and lexical decision responses for more than 25,000 traditional Chinese two-character words. In the current study, we expand their database by collecting norms for speeded naming reaction times (RTs) and accuracy rates, and compiling more lexical variables (e.g., phonological consistency and semantic neighborhood size). Following Tse et al.'s procedure, about 33 college-aged native Cantonese speakers in Hong Kong read aloud each word. We conducted item-level regression analyses to test the relative predictive power of orthographic variables (e.g., stroke count), phonological variables (e.g., phonological consistency), and semantic variables (e.g., semantic transparency) in naming performance. We also compared the effects of lexical variables on naming performance and Tse et al.'s lexical decision performance to examine the extent to which effects are task-specific or task-general. Freely accessible to the research community, this resource provides a valuable addition to other influential mega-databases, such as the English Lexicon Project (Balota et al., 2004Journal of Experimental Psychology: General, 133, 283-316), and furthers our understanding of Chinese word recognition processes.

Peekbank: An open, large-scale repository for developmental eye-tracking data of children's word recognition.

The ability to rapidly recognize words and link them to referents is central to children's early language development. This ability, often called word recognition in the developmental literature, is typically studied in the looking-while-listening paradigm, which measures infants' fixation on a target object (vs. a distractor) after hearing a target label. We present a large-scale, open database of infant and toddler eye-tracking data from looking-while-listening tasks. The goal of this effort is to address theoretical and methodological challenges in measuring vocabulary development. We first present how we created the database, its features and structure, and associated tools for processing and accessing infant eye-tracking datasets. Using these tools, we then work through two illustrative examples to show how researchers can use Peekbank to interrogate theoretical and methodological questions about children's developing word recognition ability.

Special issue on advanced deep learning methods for large scale repositories

Special issue on advanced deep learning methods for large scale repositories

Improved Eagle Strategy Algorithm for Dynamic Web Service Composition in the IoT: A Conceptual Approach

The Internet of Things (IoT) is now expanding and becoming more popular in most industries, which leads to vast growth in cloud computing. The architecture of IoT is integrated with cloud computing through web services. Recently, Dynamic Web Service Composition (DWSC) has been implemented to fulfill the IoT and business processes. In recent years, the number of cloud services has multiplied, resulting in cloud services providing similar services with similar functionality but varying in Quality of Services (QoS), for instance, on the response time of web services; however, existing methods are insufficient in solving large-scale repository issues. Bio-inspired algorithm methods have shown better performance in solving the large-scale service composition problems, unlike deterministic algorithms, which are restricted. Thus, an improved eagle strategy algorithm method is proposed to increase the performance that directly indicates an improvement in computation time in large-scale DWSC in a cloud-based platform and on both functional and non-functional attributes of services. By proposing the improved bio-inspired method, the computation time can be improved, especially in a large-scale repository of IoT.

Improving data quality in large-scale repositories through conflict resolution

Digital repositories rely on technical metadata to manage their objects. The output of characterization tools is aggregated and analyzed through content profiling. The accuracy and correctness of characterization tools vary; they frequently produce contradicting outputs, resulting in metadata conflicts. The resulting metadata conflicts limit scalable preservation risk assessment and repository management. This article presents and evaluates a rule-based approach to improving data quality in this scenario through expert-conducted conflict resolution. We characterize the data quality challenges and present a method for developing conflict resolution rules to improve data quality. We evaluate the method and the resulting data quality improvements in an experiment on a publicly available document collection. The results demonstrate that our approach enables the effective resolution of conflicts by producing rules that reduce the number of conflicts in the data set from 17 to 3%. This replicable method for presents a significant improvement in content profiling technology for digital repositories, since the enhanced data quality can improve risk assessment and preservation management in digital repository systems.

3D-FUTURE: 3D Furniture Shape with TextURE

The 3D CAD shapes in current 3D benchmarks are mostly collected from online model repositories. Thus, they typically have insufficient geometric details and less informative textures, making them less attractive for comprehensive and subtle research in areas such as high-quality 3D mesh and texture recovery. This paper presents 3D Furniture shape with TextURE (3D-FUTURE): a richly-annotated and large-scale repository of 3D furniture shapes in the household scenario. At the time of this technical report, 3D-FUTURE contains 9992 modern 3D furniture shapes with high-resolution textures and detailed attributes. To support the studies of 3D modeling from images, we couple the CAD models with 20,240 scene images. The room scenes are designed by professional designers or generated by an industrial scene creating system. Given the well-organized 3D-FUTURE and its characteristics, we provide a package of baseline experiments, such as joint 2D instance segmentation and 3D object pose estimation, image-based 3D shape retrieval, 3D object reconstruction from a single image, texture recovery for 3D shapes, and furniture composition, to facilitate related future researches on our database.

Detection and Classification of Cross-language Code Clone Types by Filtering the Nodes of ANTLR-generated Parse Tree

A complete and accurate cross-language clone detection tool can support software forking process that reuses the more reliable algorithms of legacy systems from one language code base to other. Cross-language clone detection also helps in building code recommendation system. This paper proposes a new technique to detect and classify cross-language clones of C and C++ programs by filtering the nodes of ANTLR-generated parse tree using a common grammar file, CPP14.g4. Parsing the input files using CPP14.g4 provides all the lexical and semantic information of input source code. Selective filtering of nodes performs serialization of two parse trees. Vector representation using term frequency inverse document frequency (TF-IDF) of the resultant tree is given as an input to cosine similarity to classify the clone types. Filtered parse tree of C and C++ increases the precision from 51% to 61%, and matching based on renaming the input/output expressions provides average precision of 91.97% and 95.37% for small scale and large scale repositories respectively. The proposed cross-language clone detection exhibits the highest precision of 95.37% in finding all types of clones (1, 2, 3 and 4) for 16,032 semantically similar clone pairs of C and CPP codes.

Co-attentive representation learning for web services classification

The rapid adoption of services-related technologies, such as cloud computing, has lead to the explosive growth of web services. Automated service classification that groups web services by similar functionality is a widely used technique to facilitate the management and discovery of web services within a large-scale repository. The existing service classification approaches primarily focus on learning the isolated representations of service features but ignored their internal semantic correlations. To address the aforementioned issue, we propose a novel deep neural network with the Co-Attentive Representation Learning (CARL-Net) mechanism for effectively classifying services by learning interdependent characteristics of service without feature engineering. Specifically, we propose a service data augmentation mechanism by extracting informative words from the service description using information gain theory. Such a mechanism can learn a correlation matrix among embedded augmented data and description, thereby obtaining their interdependent semantic correlation representations for service classification. We evaluate the effectiveness of CARL-Net by comprehensive experiments based on a real-world dataset collected from ProgrammableWeb, which includes 10,943 web services. Compared with seven web service classification baselines based on CNN, LSTM, Recurrent-CNN, C-LSTM, BLSTM, ServeNet and ServeNet-BERT, the CARL-Net can achieve an improvement of 5.66%–172.21% in the F-measure of web service classification.

Recommendation of technological profiles to collaborate in software projects using document embeddings

The information technology sector is continuously growing, and there is a high demand for developers. In the area of software development projects, fixing bugs or solving issues is a task that could be optimized to improve the productivity of developers. Making an adequate allocation for bug fixing will save overall project development time. Moreover, the problem will last for the shortest possible time, minimizing any negative impacts in case the project is already in production. This research work’s objective is to identify the most apt users (where the term “user” refers to any technology professional, for example a software developer, who has registered on any given platform), from a set of different user profiles, for fixing bugs in a software project. The study has been carried out by analyzing large-scale repositories of open-source projects with a large historical volume of bugs, and the extracted knowledge has been successfully applied to new, unrelated projects. Different similarity-based profile raking procedures have been studied, including neural-network-based incidence representation. The obtained results show that the system can be directly applied to different environments and that the selected user profiles are very close to those selected by human experts, which demonstrates the correct functioning of the proposed system.

Integrating the Social Media Engine with Large-scale Open Access Repositories: A Discussion

The Social Media Engine relies on interactive computer-mediated technologies and the increased impact, readership, and alt-metrics present in open access repositories—while fostering public engagement, open social scholarship, and social knowledge creation by matching readers with publications. In this paper I focus on a discussion that explores the possibilities of integrating a search engine that ranks its results according to trends in social media with large-scale open access repositories. Ultimately, this discussion aims to explore the implications of creating tools to emphasize the connections between documents that can be treated as objects of study as well.

Universal Southern blot protocol with cold or radioactive probes for the validation of alleles obtained by homologous recombination.

The widespread availability of recombineered vectors and gene targeted embryonic stem cells from large-scale repositories facilitates the generation of mouse models for functional genetic studies. Southern blotting validates the structure of these targeted alleles produced by homologous recombination, as well as indicating any additional integrations of the vector into the genome. Traditionally this technique employs radioactively-labelled probes; however, there are many laboratories that are restricted in their use of radioactivity. Here, we present a widely applicable protocol for Southern blot analysis using cold probes and alternative procedures employing radioactive probes. Furthermore, the probes are designed to recognise standardised regions of gene-targeting cassettes and so represent universally applicable reagents for assessing allelic integrity.

Biobanking in Psychiatry

Medical biobanking is concerned with establishing and maintaining large-scale repositories of biological specimens combined with comprehensive archives of clinical and biographical information on donors. This aims for controlled high and consistent quality of specimens for future biomedical research. One major objective is to assemble multiple blood components for various types of biochemical analysis and experimentation including different isolated cell types. With proper cryo-conservation, blood-derived cells can be conserved and revitalized after thawing and employed as in-vitro cell models carrying specific biological traits of donors. Optimizing pre-analytical methods can reduce pre-analytical variance thereby reducing imprecision of analytical data. This is particularly valuable for multivariate analyses of biological systems ("omics") and biomarker research. Introducing biobanking to psychiatry carries the challenge of making diagnostic allocation more compatible with biological entities than is achieved with current diagnostic categories of ICD-10 or DSM-V. Diagnostic or transdiagnostic subgroups can be stratified using biologically anchored clinical criteria. An important ethical issue of biobanking is the need for broad consent by the donors for specimen use in not yet defined future research projects. The organizational, logistic and financial costs of establishing and maintaining a biobank are considerable, but seem well warranted in view of the gainable advances in biomedical research quality, translations and clinical applications.

Ontology-driven integrative analysis of omics data through Onassis

Public repositories of large-scale omics datasets represent a valuable resource for researchers. In fact, data re-analysis can either answer novel questions or provide critical data able to complement in-house experiments. However, despite the development of standards for the compilation of metadata, the identification and organization of samples still constitutes a major bottleneck hampering data reuse. We introduce Onassis, an R package within the Bioconductor environment providing key functionalities of Natural Language Processing (NLP) tools. Leveraging biomedical ontologies, Onassis greatly simplifies the association of samples from large-scale repositories to their representation in terms of ontology-based annotations. Moreover, through the use of semantic similarity measures, Onassis hierarchically organizes the datasets of interest, thus supporting the semantically aware analysis of the corresponding omics data. In conclusion, Onassis leverages NLP techniques, biomedical ontologies, and the R statistical framework, to identify, relate, and analyze datasets from public repositories. The tool was tested on various large-scale datasets, including compendia of gene expression, histone marks, and DNA methylation, illustrating how it can facilitate the integrative analysis of various omics data.

A hybrid collaborative filtering recommendation algorithm for requirements elicitation

Requirements elicitation is one of the most critical and difficult tasks in software development. Requirements reuse has shown to be an effective and efficient elicitation technique that can enhance the quality of the requirements elicitation process and, as a result, lead to a project's success. However, the information overload problem, which is caused by the rapidly growing number of reusable software requirements in large requirements repositories, hinders the effectiveness of the requirements reuse process. Recommender systems proved to be a well-known solution to such problems. This paper focuses on the adoption of recommender systems to mitigate the problem of information overload that is inherent in the requirements elicitation process, specifically by assisting requirement engineers in retrieving relevant reusable requirements from large-scale requirements repositories. The validation results on the RALIC dataset illustrate that the proposed algorithm outperforms and mitigates the drawbacks of the benchmark collaborative filtering-based recommendation approaches.