To support physicians in clinical decision process on patients affected by Coronavirus Disease 2019 (COVID-19) in areas with a low vaccination rate, we devised and evaluated the performances of several machine learning (ML) classifiers fed with readily available clinical and laboratory data. Our observational retrospective study collected data from a cohort of 779 COVID-19 patients presenting to three hospitals of the Lazio-Abruzzo area (Italy). Based on a different selection of clinical and respiratory (ROX index and PaO2/FiO2 ratio) variables, we devised an AI-driven tool to predict safe discharge from ED, disease severity and mortality during hospitalization. To predict safe discharge our best classifier is an RF integrated with ROX index that reached AUC of 0.96. To predict disease severity the best classifier was an RF integrated with ROX index that reached an AUC of 0.91. For mortality prediction the best classifier was an RF integrated with ROX index, that reached an AUC of 0.91. The results obtained thanks to our algorithms are consistent with the scientific literature an accomplish significant performances to forecast safe discharge from ED and severe clinical course of COVID-19.

Nicotinamide adenine dinucleotide (NAD) levels are essential for the normal physiology of the cell and are strictly regulated to prevent pathological conditions. NAD functions as a coenzyme in redox reactions, as a substrate of regulatory proteins, and as a mediator of protein-protein interactions. The main objectives of this study were to identify the NAD-binding and NAD-interacting proteins, and to uncover novel proteins and functions that could be regulated by this metabolite. It was considered if cancer-associated proteins were potential therapeutic targets. Using multiple experimental databases, we defined datasets of proteins that directly interact with NAD - the NAD-binding proteins (NADBPs) dataset - and of proteins that interact with NADBPs - the NAD-protein-protein interactions (NAD-PPIs) dataset. Pathway enrichment analysis revealed that NADBPs participate in several metabolic pathways, while NAD-PPIs are mostly involved in signalling pathways. These include disease-related pathways, namely, three major neurodegenerative disorders: Alzheimer's disease, Huntington's disease, and Parkinson's disease. Then, the complete human proteome was further analysed to select potential NADBPs. TRPC3 and isoforms of diacylglycerol (DAG) kinases, which are involved in calcium signalling, were identified as new NADBPs. Potential therapeutic targets that interact with NAD were identified, that have regulatory and signalling functions in cancer and neurodegenerative diseases.

This special issue of the Journal of Integrative Bioinformatics contains updated specifications of COMBINE standards in systems and synthetic biology. The 2022 special issue presents three updates to the standards: CellML 2.0.1, SBML Level 3 Package: Spatial Processes, Version 1, Release 1, and Synthetic Biology Open Language (SBOL) Version 3.1.0. This document can also be used to identify the latest specifications for all COMBINE standards. In addition, this editorial provides a brief overview of the COMBINE 2022 meeting in Berlin.

Analyzing metabolic pathways in systems biology requires accurate kinetic parameters that represent the simulated in vivo processes. Simulation of the fermentation pathway in the Saccharomyces cerevisiae kinetic model help saves much time in the optimization process. Fitting the simulated model into the experimental data is categorized under the parameter estimation problem. Parameter estimation is conducted to obtain the optimal values for parameters related to the fermentation process. This step is essential because insufficient identification of model parameters can cause erroneous conclusions. The kinetic parameters cannot be measured directly. Therefore, they must be estimated from the experimental data either in vitro or in vivo. Parameter estimation is a challenging task in the biological process due to the complexity and nonlinearity of the model. Therefore, we propose the Artificial Bee Colony algorithm (ABC) to estimate the parameters in the fermentation pathway of S. cerevisiae to obtain more accurate values. A metabolite with a total of six parameters is involved in this article. The experimental results show that ABC outperforms other estimation algorithms and gives more accurate kinetic parameter values for the simulated model. Most of the estimated kinetic parameter values obtained from the proposed algorithm are the closest to the experimental data.

NDM-1 (New-Delhi-Metallo-β-lactamase-1) is an enzyme developed by bacteria that is implicated in bacteria resistance to almost all known antibiotics. In this study, we deliver a new, curated NDM-1 bioactivities database, along with a set of unifying rules for managing different activity properties and inconsistencies. We define the activity classification problem in terms of Multiple Instance Learning, employing embeddings corresponding to molecular substructures and present an ensemble ranking and classification framework, relaying on a k-fold Cross Validation method employing a per fold hyper-parameter optimization procedure, showing promising generalization ability. The MIL paradigm displayed an improvement up to 45.7 %, in terms of Balanced Accuracy, in comparison to the classical Machine Learning paradigm. Moreover, we investigate different compact molecular representations, based on atomic or bi-atomic substructures. Finally, we scanned the Drugbank for strongly active compounds and we present the top-15 ranked compounds.

Transmembrane transport proteins (transporters) play a crucial role in the fundamental cellular processes of all organisms by facilitating the transport of hydrophilic substrates across hydrophobic membranes. Despite the availability of numerous membrane protein sequences, their structures and functions remain largely elusive. Recently, natural language processing (NLP) techniques have shown promise in the analysis of protein sequences. Bidirectional Encoder Representations from Transformers (BERT) is an NLP technique adapted for proteins to learn contextual embeddings of individual amino acids within a protein sequence. Our previous strategy, TooT-BERT-T, differentiated transporters from non-transporters by employing a logistic regression classifier with fine-tuned representations from ProtBERT-BFD. In this study, we expand upon this approach by utilizing representations from ProtBERT, ProtBERT-BFD, and MembraneBERT in combination with classical classifiers. Additionally, we introduce TooT-BERT-CNN-T, a novel method that fine-tunes ProtBERT-BFD and discriminates transporters using a Convolutional Neural Network (CNN). Our experimental results reveal that CNN surpasses traditional classifiers in discriminating transporters from non-transporters, achieving an MCC of 0.89 and an accuracy of 95.1 % on the independent test set. This represents an improvement of 0.03 and 1.11 percentage points compared to TooT-BERT-T, respectively.

Crop plant breeding involves selecting and developing new plant varieties with desirable traits such as increased yield, improved disease resistance, and enhanced nutritional value. With the development of high-throughput technologies, such as genomics, transcriptomics, and metabolomics, crop breeding has entered a new era. However, to effectively use these technologies, integration of multi-omics data from different databases is required. Integration of omics data provides a comprehensive understanding of the biological processes underlying plant traits and their interactions. This review highlights the importance of integrating omics databases in crop plant breeding, discusses available omics data and databases, describes integration challenges, and highlights recent developments and potential benefits. Taken together, the integration of omics databases is a critical step towards enhancing crop plant breeding and improving global food security.

This work aims to describe the observed enrichment of inverted repeats in the human genome; and to identify and describe, with detailed length profiles, the regions with significant and relevant enriched occurrence of inverted repeats. The enrichment is assessed and tested with a recently proposed measure (z-scores based measure). We simulate a genome using an order 7 Markov model trained with the data from the real genome. The simulated genome is used to establish the critical values which are used as decision thresholds to identify the regions with significant enriched concentrations. Several human genome regions are highly enriched in the occurrence of inverted repeats. This is observed in all the human chromosomes. The distribution of inverted repeat lengths varies along the genome. The majority of the regions with severely exaggerated enrichment contain mainly short length inverted repeats. There are also regions with regular peaks along the inverted repeats lengths distribution (periodic regularities) and other regions with exaggerated enrichment for long lengths (less frequent). However, adjacent regions tend to have similar distributions.

EvoPPI (http://evoppi.i3s.up.pt), a meta-database for protein-protein interactions (PPI), has been upgraded (EvoPPI3) to accept new types of data, namely, PPI from patients, cell lines, and animal models, as well as data from gene modifier experiments, for nine neurodegenerative polyglutamine (polyQ) diseases caused by an abnormal expansion of the polyQ tract. The integration of the different types of data allows users to easily compare them, as here shown for Ataxin-1, the polyQ protein involved in spinocerebellar ataxia type 1 (SCA1) disease. Using all available datasets and the data here obtained for Drosophila melanogaster wt and exp Ataxin-1 mutants (also available at EvoPPI3), we show that, in humans, the Ataxin-1 network is much larger than previously thought (380 interactors), with at least 909 interactors. The functional profiling of the newly identified interactors is similar to the ones already reported in the main PPI databases. 16 out of 909 interactors are putative novel SCA1 therapeutic targets, and all but one are already being studied in the context of this disease. The 16 proteins are mainly involved in binding and catalytic activity (mainly kinase activity), functional features already thought to be important in the SCA1 disease.

Diagnosing diabetes early is critical as it helps patients live with the disease in a healthy way - through healthy eating, taking appropriate medical doses, and making patients more vigilant in their movements/activities to avoid wounds that are difficult to heal for diabetic patients. Data mining techniques are typically used to detect diabetes with high confidence to avoid misdiagnoses with other chronic diseases whose symptoms are similar to diabetes. Hidden Naïve Bayes is one of the algorithms for classification, which works under a data-mining model based on the assumption of conditional independence of the traditional Naïve Bayes. The results from this research study, which was conducted on the Pima Indian Diabetes (PID) dataset collection, show that the prediction accuracy of the HNB classifier achieved 82%. As a result, the discretization method increases the performance and accuracy of the HNB classifier.

Science has become a highly competitive undertaking concerning, for example, resources, positions, students, and publications. At the same time, the number of journals presenting scientific findings skyrockets while the knowledge increase per manuscript seems to be diminishing. Science has also become ever more dependent on computational analyses. For example, virtually all biomedical applications involve computational data analysis. The science community develops many computational tools, and there are numerous alternatives for many computational tasks. The same is true for workflow management systems, leading to a tremendous duplication of efforts. Software quality is often of low concern, and typically, a small dataset is used as a proof of principle to support rapid publication. Installation and usage of such tools are complicated, so virtual machine images, containers, and package managers are employed more frequently. These simplify installation and ease of use but do not solve the software quality issue and duplication of effort. We believe that a community-wide collaboration is needed to (a) ensure software quality, (b) increase reuse of code, (c) force proper software review, (c) increase testing, and (d) make interoperability more seamless. Such a science software ecosystem will overcome current issues and increase trust in current data analyses.

The differentiation of regions with coding potential from non-coding regions remains a key task in computational biology. Methods such as RNAcode that exploit patterns of sequence conservation for this task have a substantial advantage in classification accuracy in particular for short coding sequences, compared to methods that rely on a single input sequence. However, they require sequence alignments as input. Frequently, suitable multiple sequence alignments are not readily available and are tedious, and sometimes difficult to construct. We therefore introduce here a new web service that provides access to the well-known coding sequence detector RNAcode with minimal user overhead. It requires as input only a single target nucleotide sequence. The service automates the collection, selection, and preparation of homologous sequences from the NCBI database, as well as the construction of the multiple sequence alignment that are needed as input for RNAcode. The service automatizes the entire pre- and postprocessing and thus makes the investigation of specific genomic regions for previously unannotated coding regions, such as small peptides or additional introns, a simple task that is easily accessible to non-expert users. RNAcode_Web is accessible online at rnacode.bioinf.uni-leipzig.de.

With the rapid growth of massively parallel sequencing technologies, still more laboratories are utilising sequenced DNA fragments for genomic analyses. Interpretation of sequencing data is, however, strongly dependent on bioinformatics processing, which is often too demanding for clinicians and researchers without a computational background. Another problem represents the reproducibility of computational analyses across separated computational centres with inconsistent versions of installed libraries and bioinformatics tools. We propose an easily extensible set of computational pipelines, called SnakeLines, for processing sequencing reads; including mapping, assembly, variant calling, viral identification, transcriptomics, and metagenomics analysis. Individual steps of an analysis, along with methods and their parameters can be readily modified in a single configuration file. Provided pipelines are embedded in virtual environments that ensure isolation of required resources from the host operating system, rapid deployment, and reproducibility of analysis across different Unix-based platforms. SnakeLines is a powerful framework for the automation of bioinformatics analyses, with emphasis on a simple set-up, modifications, extensibility, and reproducibility. The framework is already routinely used in various research projects and their applications, especially in the Slovak national surveillance of SARS-CoV-2.

Core facilities have to offer technologies that best serve the needs of their users and provide them a competitive advantage in research. They have to set up and maintain instruments in the range of ten to a hundred, which produce large amounts of data and serve thousands of active projects and customers. Particular emphasis has to be given to the reproducibility of the results. More and more, the entire process from building the research hypothesis, conducting the experiments, doing the measurements, through the data explorations and analysis is solely driven by very few experts in various scientific fields. Still, the ability to perform the entire data exploration in real-time on a personal computer is often hampered by the heterogeneity of software, the data structure formats of the output, and the enormous data sizes. These impact the design and architecture of the implemented software stack. At the Functional Genomics Center Zurich (FGCZ), a joint state-of-the-art research and training facility of ETH Zurich and the University of Zurich, we have developed the B-Fabric system, which has served for more than a decade, an entire life sciences community with fundamental data science support. In this paper, we sketch how such a system can be used to glue together data (including metadata), computing infrastructures (clusters and clouds), and visualization software to support instant data exploration and visual analysis. We illustrate our in-daily life implemented approach using visualization applications of mass spectrometry data.

Over the last years it has been observed that the progress in data collection in life science has created increasing demand and opportunities for advanced bioinformatics. This includes data management as well as the individual data analysis and often covers the entire data life cycle. A variety of tools have been developed to store, share, or reuse the data produced in the different domains such as genotyping. Especially imputation, as a subfield of genotyping, requires good Research Data Management (RDM) strategies to enable use and re-use of genotypic data. To aim for sustainable software, it is necessary to develop tools and surrounding ecosystems, which are reusable and maintainable. Reusability in the context of streamlined tools can e.g. be achieved by standardizing the input and output of the different tools and adapting to open and broadly used file formats. By using such established file formats, the tools can also be connected with others, improving the overall interoperability of the software. Finally, it is important to build strong communities that maintain the tools by developing and contributing new features and maintenance updates. In this article, concepts for this will be presented for an imputation service.

Genome-scale metabolic models (GEMs) are essential tools for in silico phenotype prediction and strain optimisation. The most straightforward GEMs reconstruction approach uses published models as templates to generate the initial draft, requiring further curation. Such an approach is used by BiGG Integration Tool (BIT), available for merlin users. This tool uses models from BiGG Models database as templates for the draft models. Moreover, BIT allows the selection between different template combinations. The main objective of this study is to assess the draft models generated using this tool and compare them BIT, comparing these to CarveMe models, both of which use the BiGG database, and curated models. For this, three organisms were selected, namely Streptococcus thermophilus, Xylella fastidiosa and Mycobacterium tuberculosis. The models' variability was assessed using reactions and genes' metabolic functions. This study concluded that models generated with BIT for each organism were differentiated, despite sharing a significant portion of metabolic functions. Furthermore, the template seems to influence the content of the models, though to a lower extent. When comparing each draft with curated models, BIT had better performances than CarveMe in all metrics. Hence, BIT can be considered a fast and reliable alternative for draft reconstruction for bacteria models.

The implementation of Artificial Intelligence (AI) still faces significant hurdles and one key factor is the access to data. One approach that could support that is federated machine learning (FL) since it allows for privacy preserving data access. For this proof of concept, a prediction model for coronary artery calcification scores (CACS) has been applied. The FL was trained based on the data in the different institutions, while the centralized machine learning model was trained on one allocation of data. Both algorithms predict patients with risk scores ≥5 based on age, biological sex, waist circumference, dyslipidemia and HbA1c. The centralized model yields a sensitivity of c. 66% and a specificity of c. 70%. The FL slightly outperforms that with a sensitivity of 67% while slightly underperforming it with a specificity of 69%. It could be demonstrated that CACS prediction is feasible via both, a centralized and an FL approach, and that both show very comparable accuracy. In order to increase accuracy, additional and a higher volume of patient data is required and for that FL is utterly necessary. The developed "CACulator" serves as proof of concept, is available as research tool and shall support future research to facilitate AI implementation.

Parkinson's disease is considered the second most frequent neurodegenerative disease. It is described by the loss of dopaminergic neurons in the mid-brain. For many decades, L-DOPA has been considered as the gold standard for treating Parkinson's disease motor symptoms, however, due to the decrease of efficacy, in the long run, there is an urgent need for novel antiparkinsonian drugs. Caffeine derivatives have been reported several times for their neuroprotective properties and dual blockade of monoamine oxidase (MAO) and adenosine A2A receptors (AA2AR). Natural products are currently attracting more focus due to structural diversity and safety in contrast to synthetic drugs. In the present work, computational studies were conducted on natural product-like caffeine derivatives to search for novel potent candidates acting as dual MAO-B inhibitors/AA2AR antagonists for Parkinson's disease. Our findings revealed two natural products among the top hits: CNP0202316 and CNP0365210 fulfill the requirements of drugs acting on the brain. The selected lead compounds were further studied using molecular dynamics simulation to assess their stability with MAO-B. Current findings might shift the interest towards natural-based compounds and could be exploited to further optimize caffeine derivatives into a successful dual-target-directed drug for managing and halting the neuronal damage in Parkinson's disease patients.

This paper describes the design philosophy for our cloud-based virtual reality (VR) co-creation environment (CCE) for molecular modeling. Using interactive VR simulation can provide enhanced perspectives in molecular modeling for intuitive live demonstration and experimentation in the CCE. Then the use of the CCE can enhance knowledge creation by bringing people together to share and create ideas or knowledge that may not emerge otherwise. Our prototype CCE discussed here, which was developed to demonstrate our design philosophy, has already enabled multiple members to log in and touch virtual molecules running on a cloud server with no noticeable network latency via real-time artificial intelligence techniques. The CCE plays an essential role in the rational design of molecular robot parts, which consist of bio-molecules such as DNA and protein molecules.

Biomolecular networks, including genome-scale metabolic models (GSMMs), assemble the knowledge regarding the biological processes that happen inside specific organisms in a way that allows for analysis, simulation, and exploration. With the increasing availability of genome annotations and the development of powerful reconstruction tools, biomolecular networks continue to grow ever larger. While visual exploration can facilitate the understanding of such networks, the network sizes represent a major challenge for current visualisation systems. Building on promising results from the area of immersive analytics, which among others deals with the potential of immersive visualisation for data analysis, we present a concept for a hybrid user interface that combines a classical desktop environment with a virtual reality environment for the visual exploration of large biomolecular networks and corresponding data. We present system requirements and design considerations, describe a resulting concept, an envisioned technical realisation, and a systems biology usage scenario. Finally, we discuss remaining challenges.