Meniscal lesion detection and characterization in adult knee MRI: A deep learning model approach with external validation

PurposeEvaluation of a deep learning approach for the detection of meniscal tears and their characterization (presence/absence of migrated meniscal fragment).MethodsA large annotated adult knee MRI database was built combining medical expertise of radiologists and data scientists’ tools. Coronal and sagittal proton density fat suppressed-weighted images of 11,353 knee MRI examinations (10,401 individual patients) paired with their standardized structured reports were retrospectively collected. After database curation, deep learning models were trained and validated on a subset of 8058 examinations. Algorithm performance was evaluated on a test set of 299 examinations reviewed by 5 musculoskeletal specialists and compared to general radiologists’ reports. External validation was performed using the publicly available MRNet database. Receiver Operating Characteristic (ROC) curves results and Area Under the Curve (AUC) values were obtained on internal and external databases.ResultsA combined architecture of meniscal localization and lesion classification 3D convolutional neural networks reached AUC values of 0.93 (95% CI 0.82, 0.95) for medial and 0.84 (95% CI 0.78, 0.89) for lateral meniscal tear detection, and 0.91 (95% CI 0.87, 0.94) for medial and 0.95 (95% CI 0.92, 0.97) for lateral meniscal tear migration detection. External validation of the combined medial and lateral meniscal tear detection models resulted in an AUC of 0.83 (95% CI 0.75, 0.90) without further training and 0.89 (95% CI 0.82, 0.95) with fine tuning.ConclusionOur deep learning algorithm demonstrated high performance in knee menisci lesion detection and characterization, validated on an external database.     

Chemical language models for de novo drug design: Challenges and opportunities

Generative deep learning is accelerating de novo drug design, by allowing the generation of molecules with desired properties on demand. Chemical language models – which generate new molecules in the form of strings using deep learning – have been particularly successful in this endeavour. Thanks to advances in natural language processing methods and interdisciplinary collaborations, chemical language models are expected to become increasingly relevant in drug discovery. This minireview provides an overview of the current state-of-the-art of chemical language models for de novo design, and analyses current limitations, challenges, and advantages. Finally, a perspective on future opportunities is provided.     

PDBx/mmCIF Ecosystem: Foundational Semantic Tools for Structural Biology

PDBx/mmCIF, Protein Data Bank Exchange (PDBx) macromolecular Crystallographic Information Framework (mmCIF), has become the data standard for structural biology. With its early roots in the domain of small-molecule crystallography, PDBx/mmCIF provides an extensible data representation that is used for deposition, archiving, remediation, and public dissemination of experimentally determined three-dimensional (3D) structures of biological macromolecules by the Worldwide Protein Data Bank (wwPDB, wwpdb.org). Extensions of PDBx/mmCIF are similarly used for computed structure models by ModelArchive (modelarchive.org), integrative/hybrid structures by PDB-Dev (pdb-dev.wwpdb.org), small angle scattering data by Small Angle Scattering Biological Data Bank SASBDB (sasbdb.org), and for models computed generated with the AlphaFold 2.0 deep learning software suite (alphafold.ebi.ac.uk). Community-driven development of PDBx/mmCIF spans three decades, involving contributions from researchers, software and methods developers in structural sciences, data repository providers, scientific publishers, and professional societies. Having a semantically rich and extensible data framework for representing a wide range of structural biology experimental and computational results, combined with expertly curated 3D biostructure data sets in public repositories, accelerates the pace of scientific discovery. Herein, we describe the architecture of the PDBx/mmCIF data standard, tools used to maintain representations of the data standard, governance, and processes by which data content standards are extended, plus community tools/software libraries available for processing and checking the integrity of PDBx/mmCIF data. Use cases exemplify how the members of the Worldwide Protein Data Bank have used PDBx/mmCIF as the foundation for its pipeline for delivering Findable, Accessible, Interoperable, and Reusable (FAIR) data to many millions of users worldwide.     

Fine Sampling of Sequence Space for Membrane Protein Structural Biology

We describe an enhancement of traditional genomics-based approaches to improve the success of structure determination of membrane proteins. Following a broad screen of sequence space to identify initial expression-positive targets, we employ a second step to select orthologs with closely related sequences to these hits. We demonstrate that a greater percentage of these latter targets express well and are stable in detergent, increasing the likelihood of identifying candidates that will ultimately yield structural information.     

Detection and monitoring of HBV-related hepatocellular carcinoma from plasma cfDNA fragmentation profiles

Most hepatocellular carcinomas (HCCs) are associated with hepatitis B virus infection (HBV) in China. Early detection of HCC can significantly improve prognosis but is not yet fully clinically feasible. This study aims to develop methods for detecting HCC and studying the carcinogenesis of HBV using plasma cell-free DNA (cfDNA) whole-genome sequencing (WGS) data. Low coverage WGS was performed for 452 participants, including healthy individuals, hepatitis B patients, cirrhosis patients, and HCC patients. Then the sequencing data were processed using various machine learning models based on cfDNA fragmentation profiles for cancer detection. Our best model achieved a sensitivity of 87.10% and a specificity of 88.37%, and it showed an increased sensitivity with higher BCLC stages of HCC. Overall, this study proves the potential of a non-invasive assay based on cfDNA fragmentation profiles for the detection and prognosis of HCC and provides preliminary data on the carcinogenic mechanism of HBV.     

Logit model in prospective coronary heart disease (CHD) risk factors prediction in Saudi population

Analysis through logistic regression explored to investigate the relationship between binary or multivariable ordinal response probability and in one or more explanatory variables. The main objectives of this study to investigate advanced prediction risk factor of Coronary Heart Disease (CHD) using a logit model. Attempts made to reduce risk factors, increase public or professional awareness. Logit model used to evaluate the probability of a person develop CHD, considering any factors such as age, gender, high low-density lipoprotein (LDL) cholesterol, low high-density lipoprotein (HDL) cholesterol, high blood pressure, family history of CHD younger than 45, diabetes, smoking, being post-menopausal for women and being older than 45 for men. Logit concept of brief statistics described with slight modification to estimate the parameters testing for the significance of the coefficients, confidence interval fits the simple, multiple logit models. Besides, interpretation of the fitted logit regression model introduced. Variables showing best results within the scientific context, good explanation data assessed to fit an estimated logit model containing chosen variables, this present experiment used the statistical inference procedure; chi-square distribution, likelihood ratio, Score, or Wald test and goodness-of-fit. Health promotion started with increased public or professional awareness improved for early detection of CHD, to reduce the risk of mortality, aimed to be Saudi vision by 2030.     

Heat stress during development affects immunocompetence in workers,queens and drones of Africanized honey bees (Apis mellifera L.) (Hymenoptera: Apidae)

Though social insects generally seem to have a reduced individual immunoresponse compared to solitary species, the impact of heat stress on that response has not been studied. In the honey bee, the effect of heat stress on reproductives (queens and males/drones) may also vary compared to workers, but this is currently unknown. Here, we quantified the activity of an enzyme linked to the immune response in insects and known to be affected by heat stress in solitary species: phenoloxidase (PO), in workers, queens and drones of Africanized honey bees (AHBs) experimentally subjected to elevated temperatures during the pupal stage. Additionally, we evaluated this marker in individuals experimentally infected with the entomopathogenic fungus Metarhizium anisopliae. Differences in PO activity were found between sexes and castes, with PO activity generally higher in workers and lower in reproductives. Such differences are associated with the likelihood of exposure to infection and the role of different individuals in the colony. Contrary to our expectation, heat stress did not cause an increase in PO activity equally in all classes of individual. Heat stress during the pupal stage significantly decreased the PO activity of AHB queens, but not that of workers or drones, which more frequently engage in extranidal activity. Experimental infection with Metarhizium anisopliae reduced PO activity in queens and workers, but increased it in drones. Notably, heat stressed workers lived significantly shorter after infection despite exhibiting greater PO activity than queens or drones. We suggest that this discrepancy may be related to trade-offs among immune response cascades in honey bees such as between heat shock proteins and defensin peptides used in microbial defence. Our results provide evidence for complex relationships among humoral immune responses in AHBs and suggest that heat stress could result in a reduced life expectancy of individuals.     

Phylogenetic overview of Erysiphaceae based on nrDNA and MCM7 sequences

The classification system and evolutionary history of Erysiphaceae have been studied based on the results of molecular phylogenetic analyses. However, the sequence data used for these phylogenetic estimations have been limited to the nrDNA of ca., 50 taxa, and the relationships among higher taxonomic groups are not well understood. To provide a phylogenetic overview of Erysiphaceae, we performed phylogenetic estimations based on nrDNA and MCM7 sequences obtained from ca., 270 taxa. The phylogenetic tree showed a similar topology to the trees obtained in previous studies, although the branching order between Golovinomyceteae and Phyllactinieae was different and Phyllactinieae was not monophyletic. Phyllactinieae and Erysipheae were estimated to diversify after the divergence of Golovinomyceteae, suggesting an evolutionary trend in which non-catenate conidia + endoparasitic or non-catenate conidia + ectoparasitic lineages were derived from catenate conidia + ectoparasitic lineages. Phyllactinieae was divided into a clade of Phyllactinia + Leveillula and other clade(s) consisting of Pleochaeta and Queirozia. The phylogenetic hypothesis of Erysiphaceae was updated based on the largest dataset to date, but the higher-level phylogenetic relationships remain unclear. For a more robust phylogenetic hypothesis of Erysiphaceae, further sequence data, including protein coding regions, should be added to the dataset of nrDNA sequences.     

A novel deep learning method for maize disease identification based on small sample-size and complex background datasets

Maize diseases are a major source of yield loss, but due to the lack of human experience and limitations of traditional image-recognition technology, obtaining satisfactory large-scale identification results of maize diseases are difficult. Fortunately, the advancement of deep learning-based technology makes it possible to automatically identify diseases. However, it still faces issues caused by small sample sizes and complex field background, which affect the accuracy of disease identification. To address these issues, a deep learning-based method was proposed for maize disease identification in this paper. DenseNet121 was used as the main extraction network and a multi-dilated-CBAM-DenseNet (MDCDenseNet) model was built by combining the multi-dilated module and convolutional block attention module (CBAM) attention mechanism. Five models of MDCDenseNet, DenseNet121, ResNet50, MobileNetV2, and NASNetMobile were compared and tested using three kinds of maize leave images from the PlantVillage dataset and field-collected at Northeast Agricultural University in China. Furthermore, auxiliary classifier generative adversarial network (ACGAN) and transfer learning were used to expand the dataset and pre-train for optimal identification results. When tested on field-collected datasets with a complex background, the MDCDenseNet model outperformed compared to these models with an accuracy of 98.84%. Therefore, it can provide a viable reference for the identification of maize leaf diseases collected from the farmland with a small sample size and complex background.     

Artificial intelligence in molecular de novo design: Integration with experiment

In this mini review, we capture the latest progress of applying artificial intelligence (AI) techniques based on deep learning architectures to molecular de novo design with a focus on integration with experimental validation. We will cover the progress and experimental validation of novel generative algorithms, the validation of QSAR models and how AI-based molecular de novo design is starting to become connected with chemistry automation. While progress has been made in the last few years, it is still early days. The experimental validations conducted thus far should be considered proof-of-principle, providing confidence that the field is moving in the right direction.     

Conventional methods and future trends in antimicrobial susceptibility testing

Antimicrobial susceptibility testing is an essential task for selecting appropriate antimicrobial agents to treat infectious diseases. Constant evolution has been observed in methods used in the diagnostic microbiology laboratories. Disc diffusion or broth microdilution are classical and conventional phenotypic methods with long turnaround time and labour-intensive but still widely practiced as gold-standard. Scientists are striving to develop innovative, novel and faster methods of antimicrobial susceptibility testing to be applicable for routine microbiological laboratory practice and research. To meet the requirements, there is an increasing trend towards automation, genotypic and micro/nano technology-based innovations. Automation in detection systems and integration of computers for online data analysis and data sharing are giant leaps towards versatile nature of automated methods currently in use. Genotypic methods detect a specific genetic marker associated with resistant phenotypes using molecular amplification techniques and genome sequencing. Microfluidics and microdroplets are recent addition in the continuous advancement of methods that show great promises with regards to safety and speed and have the prospect to identify and monitor resistance mechanisms. Although genotypic and microfluidics methods have many exciting features, however, their applications into routine clinical laboratory practice warrant extensive validation. The main impetus behind the evolution of methods in antimicrobial susceptibility testing is to shorten the overall turnaround time in obtaining the results and to enhance the ease of sample processing. This comprehensive narrative review summarises major conventional phenotypic methods and automated systems currently in use, and highlights principles of some of the emerging genotypic and micro/nanotechnology-based methods in antimicrobial susceptibility testing.     

Extending the Horizon of Homology Detection with Coevolution-based Structure Prediction

Traditional sequence analysis algorithms fail to identify distant homologies when they lie beyond a detection horizon. In this review, we discuss how co-evolution-based contact and distance prediction methods are pushing back this homology detection horizon, thereby yielding new functional insights and experimentally testable hypotheses. Based on correlated substitutions, these methods divine three-dimensional constraints among amino acids in protein sequences that were previously devoid of all annotated domains and repeats. The new algorithms discern hidden structure in an otherwise featureless sequence landscape. Their revelatory impact promises to be as profound as the use, by archaeologists, of ground-penetrating radar to discern long-hidden, subterranean structures. As examples of this, we describe how triplicated structures reflecting longin domains in MON1A-like proteins, or UVR-like repeats in DISC1, emerge from their predicted contact and distance maps. These methods also help to resolve structures that do not conform to a “beads-on-a-string” model of protein domains. In one such example, we describe CFAP298 whose ubiquitin-like domain was previously challenging to perceive owing to a large sequence insertion within it. More generally, the new algorithms permit an easier appreciation of domain families and folds whose evolution involved structural insertion or rearrangement. As we exemplify with α1-antitrypsin, coevolution-based predicted contacts may also yield insights into protein dynamics and conformational change. This new combination of structure prediction (using innovative co-evolution based methods) and homology inference (using more traditional sequence analysis approaches) shows great promise for bringing into view a sea of evolutionary relationships that had hitherto lain far beyond the horizon of homology detection.     

A structural metagenomics pipeline for examining the gut microbiome

Metagenomic sequencing data provide a rich resource from which to expand our understanding of differential protein functions involved in human health. Here, we outline a pipeline that combines microbial whole genome sequencing with protein structure data to yield a structural metagenomics-informed atlas of microbial enzyme families of interest. Visualizing metagenomics data through a structural lens facilitates downstream studies including targeted inhibition and probe-based proteomics to define at the molecular level how different enzyme orthologs impact in vivo function. Application of this pipeline to gut microbial enzymes like glucuronidases, TMA lyases, and bile salt hydrolases is expected to pinpoint their involvement in health and disease and may aid in the development of therapeutics that target specific enzymes within the microbiome.     

Microbial community composition and hydrochemistry of underexplored geothermal waters in Croatia

In Croatia, a variety of geothermal springs with a wide temperature range and varied hydrochemical conditions exist, and they may harbor different niches for the distribution of microbial communities. In this study, 19 different sites, mainly located in central and eastern Croatia, were selected for primary characterization of spring hydrochemistry and microbial community composition. Using 16S rRNA gene amplicon sequencing, it was found that the bacterial communities that dominated most geothermal waters were related to Proteobacteria and Campylobacteria, while most archaeal sequences were related to Crenarchaeota. At the genus level, the prokaryotic community was highly site-specific and was often dominated by a single genus, including sites dominated by Hydrogenophilus, Sulfuricurvum, Sulfurovum, Thiofaba and Nitrospira, while the most abundant archaeal genera were affiliated to the ammonia-oxidizing archaea, Candidatus Nitrosotenuis and Candidatus Nitrososphaera. Whereas the microbial communities were overall highly location-specific, temperature, pH, ammonia, nitrate, total nitrogen, sulfate and hydrogen sulfide, as well as dissolved organic and inorganic carbon, were the abiotic factors that significantly affected microbial community composition. Furthermore, an aquifer-type effect was observed in the community composition, but there was no pronounced seasonal variability for geothermal spring communities (i.e. the community structure was mainly stable during the three seasons sampled). These results surprisingly pointed to stable and geographically unique microbial communities that were adapted to different geothermal water environments throughout Croatia. Knowing which microbial communities are present in these extreme habitats is essential for future research. They will allow us to explore further the microbial metabolisms prevailing at these geothermal sites that have high potential for biotechnological uses, as well as the establishment of the links between microbial community structure and the physicochemical environment of geothermal waters.     

Clinical and Molecular Attributes and Evaluation of Pancreatic Cystic Neoplasm

Intraductal papillary mucinous neoplasms (IPMNs) and mucinous cystic neoplasms (MCNs) are all considered “Pancreatic cystic neoplasms (PCNs)” and show a varying risk of developing into pancreatic ductal adenocarcinoma (PDAC). These lesions display different molecular characteristics, mutations, and clinical manifestations. A lack of detailed understanding of PCN subtype characteristics and their molecular mechanisms limits the development of efficient diagnostic tools and therapeutic strategies for these lesions. Proper in vivo mouse models that mimic human PCNs are also needed to study the molecular mechanisms and for therapeutic testing. A comprehensive understanding of the current status of PCN biology, mechanisms, current diagnostic methods, and therapies will help in the early detection and proper management of patients with these lesions and PDAC. This review aims to describe all these aspects of PCNs, specifically IPMNs, by describing the future perspectives.     

Prediction of Quality-control Degradation Signals in Yeast Proteins

Effective proteome homeostasis is key to cellular and organismal survival, and cells therefore contain efficient quality control systems to monitor and remove potentially toxic misfolded proteins. Such general protein quality control to a large extent relies on the efficient and robust delivery of misfolded or unfolded proteins to the ubiquitin–proteasome system. This is achieved via recognition of so-called degradation motifs—degrons—that are assumed to become exposed as a result of protein misfolding. Despite their importance, the nature and sequence properties of quality-control degrons remain elusive. Here, we have used data from a yeast-based screen of 23,600 17-residue peptides to build a predictor of quality-control degrons. The resulting model, QCDPred (Quality Control Degron Prediction), achieves good accuracy using only the sequence composition of the peptides as input. Our analysis reveals that strong degrons are enriched in hydrophobic amino acids and depleted in negatively charged amino acids, in line with the expectation that they are buried in natively folded proteins. We applied QCDPred to the yeast proteome, enabling us to analyse more widely the potential effects of degrons. As an example, we show a correlation between cellular abundance and degron potential in disordered regions of proteins. Together with recent results on membrane proteins, our work suggest that the recognition of exposed hydrophobic residues is a key and generic mechanism for proteome homeostasis. QCDPred is freely available as open source code and via a web interface.     

TAPASS: Tool for annotation of protein amyloidogenicity in the context of other structural states

Numerous studies have demonstrated that the propensity of a protein to form amyloids or amorphous aggregates is encoded by its amino acid sequence. This led to the emergence of several computational programs to predict amyloidogenicity from amino acid sequences. However, a growing number of studies indicate that an accurate prediction of the protein aggregation can only be achieved when also accounting for the overall structural context of the protein, and the likelihood of transition between the initial state and the aggregate. Here, we describe a computational pipeline called TAPASS, which was designed to do just that. The pipeline assigns each residue of a protein as belonging to a structured region or an intrinsically disordered region (IDR). For this purpose, TAPASS uses either several state-of-the-art programs for prediction of IDRs, of transmembrane regions and of structured domains or the artificial intelligence program AlphaFold. In the next step, this assignment is crossed with amyloidogenicity prediction. As a result, TAPASS allows the detection of Exposed Amyloidogenic Regions (EARs) located within intrinsically disordered regions (IDRs) and carrying high amyloidogenic potential. TAPASS can substantially improve the prediction of amyloids and be used in proteome-wide analysis to discover new amyloid-forming proteins. Its results, combined with clinical data, can create individual risk profiles for different amyloidoses, opening up new opportunities for personalised medicine. The architecture of the pipeline is designed so that it makes it easy to add new individual predictors as they become available. TAPASS can be used through the web interface (https://bioinfo.crbm.cnrs.fr/index.php?route=tools&tool=32).     

Dynamic Protein-RNA recognition in primary MicroRNA processing

MicroRNAs are prevalent regulators of gene expression, controlling most of the proteome in multicellular organisms. To generate the functional small RNAs, precise processing steps are required. In animals, microRNA biogenesis is initiated by Microprocessor that minimally consists of the Drosha enzyme and its partner, DGCR8. This first step is critical for selecting primary microRNAs, and many RNA-binding proteins and regulatory pathways target both the accuracy and efficiency of microRNA maturation. Structures of Drosha and DGCR8 in complex with primary microRNAs elucidate how RNA structural features rather than sequence provide the framework for substrate recognition. Comparing multiple states of Microprocessor and the closely related Dicer homologs shed light on the dynamic protein-RNA complex assembly and disassembly required to recognize RNAs with diverse sequences via common structural features.