Challenges of big data integration in the life sciences

Current state of data stewardship tools in life science

In today's data-centric landscape, effective data stewardship is critical for facilitating scientific research and innovation. This article provides an overview of essential tools and frameworks for modern data stewardship practices. Over 300 tools were analyzed in this study, assessing their utility, relevance to data stewardship, and applicability within the life sciences domain.

De Novo Generation of Chemical Structures of Inhibitor and Activator Candidates for Therapeutic Target Proteins by a Transformer-Based Variational Autoencoder and Bayesian Optimization

Deep generative models for molecular generation have been gaining much attention as structure generators to accelerate drug discovery. However, most previously developed methods are chemistry-centric approaches, and comprehensive biological responses in the cell have not been taken into account. In this study, we propose a novel computational method, TRIOMPHE-BOA (transcriptome-based inference and generation of molecules with desired phenotypes using the Bayesian optimization algorithm), to generate new chemical structures of inhibitor or activator candidates for therapeutic target proteins by integrating chemically and genetically perturbed transcriptome profiles. In the algorithm, the substructures of multiple molecules that were selected based on the transcriptome analysis are fused in the design of new chemical structures by exploring the latent space of a Transformer-based variational autoencoder using Bayesian optimization. Our results demonstrate the usefulness of the proposed method in terms of having high reproducibility of existing ligands for 10 therapeutic target proteins when compared with previous methods. Moreover, this method can be applied to proteins without detailed 3D structures or known ligands and is expected to become a powerful tool for more efficient hit identification.

MHIF-MSEA: a novel model of miRNA set enrichment analysis based on multi-source heterogeneous information fusion

Introduction: MicroRNAs (miRNAs) are a class of non-coding RNA molecules that play a crucial role in the regulation of diverse biological processes across various organisms. Despite not encoding proteins, miRNAs have been found to have significant implications in the onset and progression of complex human diseases. Methods: Conventional methods for miRNA functional enrichment analysis have certain limitations, and we proposed a novel method called MiRNA Set Enrichment Analysis based on Multi-source Heterogeneous Information Fusion (MHIF-MSEA). Three miRNA similarity networks (miRSN-DA, miRSN-GOA, and miRSN-PPI) were constructed in MHIF-MSEA. These networks were built based on miRNA-disease association, gene ontology (GO) annotation of target genes, and protein-protein interaction of target genes, respectively. These miRNA similarity networks were fused into a single similarity network with the averaging method. This fused network served as the input for the random walk with restart algorithm, which expanded the original miRNA list. Finally, MHIF-MSEA performed enrichment analysis on the expanded list. Results and Discussion: To determine the optimal network fusion approach, three case studies were introduced: colon cancer, breast cancer, and hepatocellular carcinoma. The experimental results revealed that the miRNA-miRNA association network constructed using miRSN-DA and miRSN-GOA exhibited superior performance as the input network. Furthermore, the MHIF-MSEA model performed enrichment analysis on differentially expressed miRNAs in breast cancer and hepatocellular carcinoma. The achieved p-values were 2.17e(-75) and 1.50e(-77), and the hit rates improved by 39.01% and 44.68% compared to traditional enrichment analysis methods, respectively. These results confirm that the MHIF-MSEA method enhances the identification of enriched miRNA sets by leveraging multiple sources of heterogeneous information, leading to improved insights into the functional implications of miRNAs in complex diseases.

De novo drug design based on patient gene expression profiles via deep learning

Computational de novo drug design is a challenging issue in medicine, and it is desirable to consider all of the relevant information of the biological systems in a disease state. Here, we propose a novel computational method to generate drug candidate molecular structures from patient gene expression profiles via deep learning, which we call DRAGONET. Our model can generate new molecules that are likely to counteract disease-specific gene expression patterns in patients, which is made possible by exploring the latent space constructed by a transformer-based variational autoencoder and integrating the substructures of disease-correlated molecules. We applied DRAGONET to generate drug candidate molecules for gastric cancer, atopic dermatitis, and Alzheimer's disease, and demonstrated that the newly generated molecules were chemically similar to registered drugs for each disease. This approach is applicable to diseases with unknown therapeutic target proteins and will make a significant contribution to the field of precision medicine.

Network-informed discovery of multidrug combinations for ERα+/HER2-/PI3Kα-mutant breast cancer

Breast cancer is a persistent threat to women worldwide. A large proportion of breast cancers are dependent on the estrogen receptor α (ERα) for tumor progression. Therefore, targeting ERα with antagonists, such as tamoxifen, or estrogen deprivation by aromatase inhibitors remain standard therapies for ERα + breast cancer. The clinical benefits of monotherapy are often counterbalanced by off-target toxicity and development of resistance. Combinations of more than two drugs might be of great therapeutic value to prevent resistance, and to reduce doses, and hence, decrease toxicity. We mined data from the literature and public repositories to construct a network of potential drug targets for synergistic multidrug combinations. With 9 drugs, we performed a phenotypic combinatorial screen with ERα + breast cancer cell lines. We identified two optimized low-dose combinations of 3 and 4 drugs of high therapeutic relevance to the frequent ERα + /HER2-/PI3Kα-mutant subtype of breast cancer. The 3-drug combination targets ERα in combination with PI3Kα and cyclin-dependent kinase inhibitor 1 (p21). In addition, the 4-drug combination contains an inhibitor for poly (ADP-ribose) polymerase 1 (PARP1), which showed benefits in long-term treatments. Moreover, we validated the efficacy of the combinations in tamoxifen-resistant cell lines, patient-derived organoids, and xenograft experiments. Thus, we propose multidrug combinations that have the potential to overcome the standard issues of current monotherapies.

TLSEA: a tool for lncRNA set enrichment analysis based on multi-source heterogeneous information fusion

Long non-coding RNAs (lncRNAs) play an important regulatory role in gene transcription and post-transcriptional modification, and lncRNA regulatory dysfunction leads to a variety of complex human diseases. Hence, it might be beneficial to detect the underlying biological pathways and functional categories of genes that encode lncRNA. This can be carried out by using gene set enrichment analysis, which is a pervasive bioinformatic technique that has been widely used. However, accurately performing gene set enrichment analysis of lncRNAs remains a challenge. Most conventional enrichment analysis methods have not exhaustively included the rich association information among genes, which usually affects the regulatory functions of genes. Here, we developed a novel tool for lncRNA set enrichment analysis (TLSEA) to improve the accuracy of the gene functional enrichment analysis, which extracted the low-dimensional vectors of lncRNAs in two functional annotation networks with the graph representation learning method. A novel lncRNA-lncRNA association network was constructed by merging lncRNA-related heterogeneous information obtained from multiple sources with the different lncRNA-related similarity networks. In addition, the random walk with restart method was adopted to effectively expand the lncRNAs submitted by users according to the lncRNA-lncRNA association network of TLSEA. In addition, a case study of breast cancer was performed, which demonstrated that TLSEA could detect breast cancer more accurately than conventional tools. The TLSEA can be accessed freely at http://www.lirmed.com:5003/tlsea.

pISA-tree - a data management framework for life science research projects using a standardised directory tree

We developed pISA-tree, a straightforward and flexible data management solution for organisation of life science project-associated research data and metadata. pISA-tree was initiated by end-user requirements thus its strong points are practicality and low maintenance cost. It enables on-the-fly creation of enriched directory tree structure (project/Investigation/Study/Assay) based on the ISA model, in a standardised manner via consecutive batch files. Templates-based metadata is generated in parallel at each level enabling guided submission of experiment metadata. pISA-tree is complemented by two R packages, pisar and seekr. pisar facilitates integration of pISA-tree datasets into bioinformatic pipelines and generation of ISA-Tab exports. seekr enables synchronisation with the FAIRDOMHub repository. Applicability of pISA-tree was demonstrated in several national and international multi-partner projects. The system thus supports findable, accessible, interoperable and reusable (FAIR) research and is in accordance with the Open Science initiative. Source code and documentation of pISA-tree are available at https://github.com/NIB-SI/pISA-tree.

Globally Accessible Distributed Data Sharing (GADDS): a decentralized FAIR platform to facilitate data sharing in the life sciences

MOTIVATION

Technical advances have revolutionized the life sciences and researchers commonly face challenges associated with handling large amounts of heterogeneous digital data. The Findable, Accessible, Interoperable and Reusable (FAIR) principles provide a framework to support effective data management. However, implementing this framework is beyond the means of most researchers in terms of resources and expertise, requiring awareness of metadata, policies, community agreements, and other factors such as vocabularies and ontologies.

RESULTS

We have developed the Globally Accessible Distributed Data Sharing (GADDS) platform to facilitate FAIR-like data-sharing in cross-disciplinary research collaborations. The platform consists of (i) a blockchain based metadata quality control system, (ii) a private cloud-like storage system and (iii) a version control system. GADDS is built with containerized technologies, providing minimal hardware standards and easing scalability, and offers decentralized trust via transparency of metadata, facilitating data exchange and collaboration. As a use case, we provide an example implementation in engineered living material technology within the Hybrid Technology Hub at the University of Oslo.

AVAILABILITY

Demo version available at https://github.com/pavelvazquez/GADDS.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Omics Analyses: How to Navigate Through a Constant Data Deluge

Omics data are very valuable for researchers in biology, but the work required to develop a solid expertise in their analysis contrasts with the rapidity with which the omics technologies evolve. Data accumulate in public databases, and despite significant advances in bioinformatics softwares to integrate them, data analysis remains a burden for those who perform experiments. Beyond the issue of dealing with a very large number of results, we believe that working with omics data requires a change in the way scientific problems are solved. In this chapter, we explain pitfalls and tips we found during our functional genomics projects in yeasts. Our main lesson is that, if applying a protocol does not guarantee a successful project, following simple rules can help to become strategic and intentional, thus avoiding an endless drift into an ocean of possibilities.

CHEK1: a hub gene related to poor prognosis for lung adenocarcinoma

Aim: The study aims to pinpoint hub genes and investigate their functions in order to gain insightful understandings of lung adenocarcinoma (LUAD). Methods: Bioinformatic approaches were adopted to investigate genes in databases including Gene Expression Omnibus, WebGestalt, STRING and Cytoscape, GEPIA2, Oncomine, Human Protein Atlas, TIMER2.0, UALCAN, cBioPortal, TargetScanHuman, OncomiR, ENCORI, Kaplan-Meier plotter, UCSC Xena, European Molecular Biology Laboratory - European Bioinformatics Institute Single Cell Expression Atlas and CancerSEA. Results: Five hub genes were ascertained. CHEK1 was overexpressed in a range of cancers, including LUAD. Promoter methylation, amplification and miRNA regulation might trigger CHEK1 upregulation, signaling poor prognosis. CHEK1 with its coexpressed genes were enriched in the cell cycle pathway. Intratumor heterogeneity of CHEK1 expression could be observed. Cell clusters with CHEK1 expression were more prone to metastasis and epithelial-to-mesenchymal transition. Conclusion: CHEK1 might potentially act as a prognostic biomarker for LUAD.

Machine learning models for classification tasks related to drug safety

In this review, we outline the current trends in the field of machine learning-driven classification studies related to ADME (absorption, distribution, metabolism and excretion) and toxicity endpoints from the past six years (2015-2021). The study focuses only on classification models with large datasets (i.e. more than a thousand compounds). A comprehensive literature search and meta-analysis was carried out for nine different targets: hERG-mediated cardiotoxicity, blood-brain barrier penetration, permeability glycoprotein (P-gp) substrate/inhibitor, cytochrome P450 enzyme family, acute oral toxicity, mutagenicity, carcinogenicity, respiratory toxicity and irritation/corrosion. The comparison of the best classification models was targeted to reveal the differences between machine learning algorithms and modeling types, endpoint-specific performances, dataset sizes and the different validation protocols. Based on the evaluation of the data, we can say that tree-based algorithms are (still) dominating the field, with consensus modeling being an increasing trend in drug safety predictions. Although one can already find classification models with great performances to hERG-mediated cardiotoxicity and the isoenzymes of the cytochrome P450 enzyme family, these targets are still central to ADMET-related research efforts.

A global collaboration to study intimate partner violence-related head trauma: The ENIGMA consortium IPV working group

Intimate partner violence includes psychological aggression, physical violence, sexual violence, and stalking from a current or former intimate partner. Past research suggests that exposure to intimate partner violence can impact cognitive and psychological functioning, as well as neurological outcomes. These seem to be compounded in those who suffer a brain injury as a result of trauma to the head, neck or body due to physical and/or sexual violence. However, our understanding of the neurobehavioral and neurobiological effects of head trauma in this population is limited due to factors including difficulty in accessing/recruiting participants, heterogeneity of samples, and premorbid and comorbid factors that impact outcomes. Thus, the goal of the Enhancing NeuroImaging Genetics through Meta-Analysis (ENIGMA) Consortium Intimate Partner Violence Working Group is to develop a global collaboration that includes researchers, clinicians, and other key community stakeholders. Participation in the working group can include collecting harmonized data, providing data for meta- and mega-analysis across sites, or stakeholder insight on key clinical research questions, promoting safety, participant recruitment and referral to support services. Further, to facilitate the mega-analysis of data across sites within the working group, we provide suggestions for behavioral surveys, cognitive tests, neuroimaging parameters, and genetics that could be used by investigators in the early stages of study design. We anticipate that the harmonization of measures across sites within the working group prior to data collection could increase the statistical power in characterizing how intimate partner violence-related head trauma impacts long-term physical, cognitive, and psychological health.

Digitalization in microbiology - Paving the path to sustainable circular bioeconomy

The transition to a sustainable bio-based circular economy requires cutting edge technologies that ensure economic growth with environmentally responsible action. This transition will only be feasible when the opportunities of digitalization are also exploited. Digital methods and big data handling have already found their way into life sciences and generally offer huge potential in various research areas. While computational analyses of microbial metagenome data have become state of the art, the true potential of bioinformatics remains mostly untapped so far. In this article we present challenges and opportunities of digitalization including multi-omics approaches in discovering and exploiting the microbial diversity of the planet with the aim to identify robust biocatalysts for application in sustainable bioprocesses as part of the transition from a fossil-based to a bio-based circular economy. This will contribute to solving global challenges, including utilization of natural resources, food supply, health, energy and the environment.