PROTAX-GPU: a scalable probabilistic taxonomic classification system for DNA barcodes.

DNA-based identification is vital for classifying biological specimens, yet methods to quantify the uncertainty of sequence-based taxonomic assignments are scarce. Challenges arise from noisy reference databases, including mislabelled entries and missing taxa. PROTAX addresses these issues with a probabilistic approach to taxonomic classification, advancing on methods that rely solely on sequence similarity. It provides calibrated probabilistic assignments to a partially populated taxonomic hierarchy, accounting for taxa that lack references and incorrect taxonomic annotation. While effective on smaller scales, global application of PROTAX necessitates substantially larger reference libraries, a goal previously hindered by computational barriers. We introduce PROTAX-GPU, a scalable algorithm capable of leveraging the global Barcode of Life Data System (>14 million specimens) as a reference database. Using graphics processing units (GPU) to accelerate similarity and nearest-neighbour operations and the JAX library for Python integration, we achieve over a 1000 × speedup compared with the central processing unit (CPU)-based implementation without compromising PROTAX's key benefits. PROTAX-GPU marks a significant stride towards real-time DNA barcoding, enabling quicker and more efficient species identification in environmental assessments. This capability opens up new avenues for real-time monitoring and analysis of biodiversity, advancing our ability to understand and respond to ecological dynamics. This article is part of the theme issue 'Towards a toolkit for global insect biodiversity monitoring'.

Improving the Barrier and Mechanical Properties of Paper Used for Packing Applications with Renewable Hydrophobic Coatings Derived from Camelina Oil.

This study looked at using modified camelina oil to develop sustainable coatings that could replace those derived from petroleum-based materials for use in packaging and other industrial sectors. Solvent-free synthesis of maleic anhydride grafted camelina oil (MCO) was carried out at two different temperatures (200 and 230 °C) to obtain sustainable hydrophobic coating materials for paper substrates. Maleic anhydride grafting of camelina oil was confirmed with attenuated total reflectance-Fourier transform infrared and NMR spectroscopic techniques, and up to 16% grafting of maleic anhydride was achieved, as determined by the titration method. MCO, obtained at different reaction temperatures, was coated onto cellulosic paper and evaluated for its hydrophobicity, mechanical, oxygen, and water vapor barrier properties. Scanning electron microscopy indicated the homogeneous dispersion of coating material onto the paper substrate. MCO-coated papers (MCO-200C paper and MCO-230C paper) provided a water contact angle of above 90° which indicates that the modified oil was working as a hydrophobic coating. Water vapor permeability (WVP) testing of coated papers revealed a reduction in WVP of up to 94% in comparison to the uncoated paper. Moreover, an improved oxygen barrier property was also observed for paper coated with both types of MCO. Analysis of the mechanical properties showed a greater than 70% retention of tensile strength and up to a five-fold increase in elongation at break of coated versus uncoated papers. Overall, the results show that camelina oil, a renewable resource, can be modified to produce environmentally friendly hydrophobic coating materials with improved mechanical and water vapor barrier properties that can serve as a potential coating material in the packaging industry. The results of this research could find applications in the huge paper packaging industries, specially in food packaging.

Degradation versus fibrillogenesis, two alternative pathways modulated by seeds and glycosaminoglycans.

The mechanism that converts native human transthyretin into amyloid fibrils in vivo is still a debated and controversial issue. Commonly, non-physiological conditions of pH, temperature, or organic solvents are used in in vitro models of fibrillogenesis of globular proteins. Transthyretin amyloid formation can be achieved under physiological conditions through a mechano-enzymatic mechanism involving specific serine proteases such as trypsin or plasmin. Here, we investigate S52P and L111M transthyretin variants, both causing a severe form of systemic amyloidosis mostly targeting the heart at a relatively young age with heterogeneous phenotype among patients. Our studies on thermodynamics show that both proteins are significantly less stable than other amyloidogenic variants. However, despite a similar thermodynamic stability, L111M variant seems to have enhanced susceptibility to cleavage and a lower tendency to form fibrils than S52P in the presence of specific proteases and biomechanical forces. Heparin strongly enhances the fibrillogenic capacity of L111M transthyretin, but has no effect on the S52P variant. Fibrillar seeds similarly affect the fibrillogenesis of both proteins, with a stronger effect on the L111M variant. According to our model of mechano-enzymatic fibrillogenesis, both full-length and truncated monomers, released after the first cleavage, can enter into fibrillogenesis or degradation pathways. Our findings show that the kinetics of the two processes can be affected by several factors, such as intrinsic amyloidogenicity due to the specific mutations, environmental factors including heparin and fibrillar seeds that significantly accelerate the fibrillogenic pathway.

The mental health impacts of the COVID-19 pandemic among individuals with depressive, anxiety, and stressor-related disorders: A scoping review.

OBJECTIVE: A scoping review of studies published in the first year of the COVID-19 pandemic focused on individuals with pre-existing symptoms of depression, anxiety, and specified stressor-related disorders, with the objective of mapping the research conducted. ELIGIBILITY CRITERIA: (1) direct study of individuals with pre-existing depressive, anxiety, and/or specified stressor-related (i.e., posttraumatic stress, acute stress) disorders/issues; (2) focus on mental health-related pandemic effects, and; (3) direct study of mental health symptoms related to depression, anxiety, or psychological distress. SOURCES OF EVIDENCE: Database-specific subject headings and natural language keywords were searched in Medline, Embase, APA PsycInfo, and Cumulative Index to Nursing & Allied Health Literature (CINAHL) up to March 3, 2021. Review of potentially relevant studies was conducted by two independent reviewers and proceeded in two stages: (1) title and abstract review, and; (2) full paper review. DATA CHARTING: Study details (i.e., location, design and methodology, sample or population, outcome measures, and key findings) were extracted from included studies by one reviewer and confirmed by the Principal Investigator. RESULTS: 66 relevant articles from 26 countries were identified. Most studies adopted a cross-sectional design and were conducted via online survey. About half relied on general population samples, with the remainder assessing special populations, primarily mental health patients. The most commonly reported pre-existing category of disorders or symptoms was depression, followed closely by anxiety. Most studies included depressive and anxiety symptoms as outcome measures and demonstrated increased vulnerability to mental health symptoms among individuals with a pre-existing mental health issue. CONCLUSION: These findings suggest that improved mental health supports are needed during the pandemic and point to future research needs, including reviews of other diagnostic categories and reviews of research published in subsequent years of the pandemic.

Getting the bugs out of AI: Advancing ecological research on arthropods through computer vision.

Deep learning for computer vision has shown promising results in the field of entomology, however, there still remains untapped potential. Deep learning performance is enabled primarily by large quantities of annotated data which, outside of rare circumstances, are limited in ecological studies. Currently, to utilize deep learning systems, ecologists undergo extensive data collection efforts, or limit their problem to niche tasks. These solutions do not scale to region agnostic models. However, there are solutions that employ data augmentation, simulators, generative models, and self-supervised learning that can supplement limited labelled data. Here, we highlight the success of deep learning for computer vision within entomology, discuss data collection efforts, provide methodologies for optimizing learning from limited annotations, and conclude with practical guidelines for how to achieve a foundation model for entomology capable of accessible automated ecological monitoring on a global scale.

Calcium Binds to Transthyretin with Low Affinity.

The plasma protein transthyretin (TTR), a transporter for thyroid hormones and retinol in plasma and cerebrospinal fluid, is responsible for the second most common type of systemic (ATTR) amyloidosis either in its wild type form or as a result of destabilizing genetic mutations that increase its aggregation propensity. The association between free calcium ions (Ca2+) and TTR is still debated, although recent work seems to suggest that calcium induces structural destabilization of TTR and promotes its aggregation at non-physiological low pH in vitro. We apply high-resolution NMR spectroscopy to investigate calcium binding to TTR showing the formation of labile interactions, which leave the native structure of TTR substantially unaltered. The effect of calcium binding on TTR-enhanced aggregation is also assessed at physiological pH through the mechano-enzymatic mechanism. Our results indicate that, even if the binding is weak, about 7% of TTR is likely to be Ca2+-bound in vivo and therefore more aggregation prone as we have shown that this interaction is able to increase the protein susceptibility to the proteolytic cleavage that leads to aggregation at physiological pH. These events, even if involving a minority of circulating TTR, may be relevant for ATTR, a pathology that takes several decades to develop.

Amyloid Formation by Globular Proteins: The Need to Narrow the Gap Between in Vitro and in Vivo Mechanisms.

The globular to fibrillar transition of proteins represents a key pathogenic event in the development of amyloid diseases. Although systemic amyloidoses share the common characteristic of amyloid deposition in the extracellular matrix, they are clinically heterogeneous as the affected organs may vary. The observation that precursors of amyloid fibrils derived from circulating globular plasma proteins led to huge efforts in trying to elucidate the structural events determining the protein metamorphosis from their globular to fibrillar state. Whereas the process of metamorphosis has inspired poets and writers from Ovid to Kafka, protein metamorphism is a more recent concept. It is an ideal metaphor in biochemistry for studying the protein folding paradigm and investigating determinants of folding dynamics. Although we have learned how to transform both normal and pathogenic globular proteins into fibrillar polymers in vitro, the events occurring in vivo, are far more complex and yet to be explained. A major gap still exists between in vivo and in vitro models of fibrillogenesis as the biological complexity of the disease in living organisms cannot be reproduced at the same extent in the test tube. Reviewing the major scientific attempts to monitor the amyloidogenic metamorphosis of globular proteins in systems of increasing complexity, from cell culture to human tissues, may help to bridge the gap between the experimental models and the actual pathological events in patients.

Federated learning and differential privacy for medical image analysis.

The artificial intelligence revolution has been spurred forward by the availability of large-scale datasets. In contrast, the paucity of large-scale medical datasets hinders the application of machine learning in healthcare. The lack of publicly available multi-centric and diverse datasets mainly stems from confidentiality and privacy concerns around sharing medical data. To demonstrate a feasible path forward in medical image imaging, we conduct a case study of applying a differentially private federated learning framework for analysis of histopathology images, the largest and perhaps most complex medical images. We study the effects of IID and non-IID distributions along with the number of healthcare providers, i.e., hospitals and clinics, and the individual dataset sizes, using The Cancer Genome Atlas (TCGA) dataset, a public repository, to simulate a distributed environment. We empirically compare the performance of private, distributed training to conventional training and demonstrate that distributed training can achieve similar performance with strong privacy guarantees. We also study the effect of different source domains for histopathology images by evaluating the performance using external validation. Our work indicates that differentially private federated learning is a viable and reliable framework for the collaborative development of machine learning models in medical image analysis.

Plasmin activity promotes amyloid deposition in a transgenic model of human transthyretin amyloidosis.

Cardiac ATTR amyloidosis, a serious but much under-diagnosed form of cardiomyopathy, is caused by deposition of amyloid fibrils derived from the plasma protein transthyretin (TTR), but its pathogenesis is poorly understood and informative in vivo models have proved elusive. Here we report the generation of a mouse model of cardiac ATTR amyloidosis with transgenic expression of human TTRS52P. The model is characterised by substantial ATTR amyloid deposits in the heart and tongue. The amyloid fibrils contain both full-length human TTR protomers and the residue 49-127 cleavage fragment which are present in ATTR amyloidosis patients. Urokinase-type plasminogen activator (uPA) and plasmin are abundant within the cardiac and lingual amyloid deposits, which contain marked serine protease activity; knockout of α2-antiplasmin, the physiological inhibitor of plasmin, enhances amyloid formation. Together, these findings indicate that cardiac ATTR amyloid deposition involves local uPA-mediated generation of plasmin and cleavage of TTR, consistent with the previously described mechano-enzymatic hypothesis for cardiac ATTR amyloid formation. This experimental model of ATTR cardiomyopathy has potential to allow further investigations of the factors that influence human ATTR amyloid deposition and the development of new treatments.

Dementia in the older population is associated with neocortex content of serum amyloid P component.

Despite many reported associations, the direct cause of neurodegeneration responsible for cognitive loss in Alzheimer's disease and some other common dementias is not known. The normal human plasma protein, serum amyloid P component, a constituent of all human fibrillar amyloid deposits and present on most neurofibrillary tangles, is cytotoxic for cerebral neurones in vitro and in experimental animals in vivo. The neocortical content of serum amyloid P component was immunoassayed in 157 subjects aged 65 or more with known dementia status at death, in the large scale, population-representative, brain donor cohort of the Cognitive Function and Ageing Study, which avoids the biases inherent in studies of predefined clinico-pathological groups. The serum amyloid P component values were significantly higher in individuals with dementia, independent of serum albumin content measured as a control for plasma in the cortex samples. The odds ratio for dementia at death in the high serum amyloid P component tertile was 5.24 (95% confidence interval 1.79-15.29) and was independent of Braak tangle stages and Thal amyloid-ß phases of neuropathological severity. The strong and specific association of higher brain content of serum amyloid P component with dementia, independent of neuropathology, is consistent with a pathogenetic role in dementia.

C. elegans feed yolk to their young in a form of primitive lactation.

The nematode Caenorhabditis elegans exhibits rapid senescence that is promoted by the insulin/IGF-1 signalling (IIS) pathway via regulated processes that are poorly understood. IIS also promotes production of yolk for egg provisioning, which in post-reproductive animals continues in an apparently futile fashion, supported by destructive repurposing of intestinal biomass that contributes to senescence. Here we show that post-reproductive mothers vent yolk which can be consumed by larvae and promotes their growth. This implies that later yolk production is not futile; instead vented yolk functions similarly to milk. Moreover, yolk venting is promoted by IIS. These findings suggest that a self-destructive, lactation-like process effects resource transfer from postreproductive C. elegans mothers to offspring, in a fashion reminiscent of semelparous organisms that reproduce in a single, suicidal burst. That this process is promoted by IIS provides insights into how and why IIS shortens lifespan in C. elegans.

Clinical ApoA-IV amyloid is associated with fibrillogenic signal sequence.

Apolipoprotein A-IV amyloidosis is an uncommon form of the disease normally resulting in renal and cardiac dysfunction. ApoA-IV amyloidosis was identified in 16 patients attending the National Amyloidosis Centre and in eight clinical samples received for histology review. Unexpectedly, proteomics identified the presence of ApoA-IV signal sequence residues (p.18-43 to p.20-43) in 16/24 trypsin-digested amyloid deposits but in only 1/266 non-ApoA-IV amyloid samples examined. These additional signal residues were also detected in the cardiac sample from the Swedish patient in which ApoA-IV amyloid was first described, and in plasma from a single cardiac ApoA-IV amyloidosis patient. The most common signal-containing peptide observed in ApoA-IV amyloid, p.20-43, and to a far lesser extent the N-terminal peptide, p.21-43, were fibrillogenic in vitro at physiological pH, generating Congo red-positive fibrils. The addition of a single signal-derived alanine residue to the N-terminus has resulted in markedly increased fibrillogenesis. If this effect translates to the mature circulating protein in vivo, then the presence of signal may result in preferential deposition as amyloid, perhaps acting as seed for the main circulating native form of the protein; it may also influence other ApoA-IV-associated pathologies. © 2021 The Authors. The Journal of Pathology published by John Wiley & Sons, Ltd. on behalf of The Pathological Society of Great Britain and Ireland.

Clinical Amyloid Typing by Proteomics: Performance Evaluation and Data Sharing Between Two Centres.

Amyloidosis is a relatively rare human disease caused by the deposition of abnormal protein fibres in the extracellular space of various tissues, impairing their normal function. Proteomic analysis of patients' biopsies, developed by Dogan and colleagues at the Mayo Clinic, has become crucial for clinical diagnosis and for identifying the amyloid type. Currently, the proteomic approach is routinely used at National Amyloidosis Centre (NAC, London, UK) and Istituto di Tecnologie Biomediche-Consiglio Nazionale delle Ricerche (ITB-CNR, Milan, Italy). Both centres are members of the European Proteomics Amyloid Network (EPAN), which was established with the aim of sharing and discussing best practice in the application of amyloid proteomics. One of the EPAN's activities was to evaluate the quality and the confidence of the results achieved using different software and algorithms for protein identification. In this paper, we report the comparison of proteomics results obtained by sharing NAC proteomics data with the ITB-CNR centre. Mass spectrometric raw data were analysed using different software platforms including Mascot, Scaffold, Proteome Discoverer, Sequest and bespoke algorithms developed for an accurate and immediate amyloid protein identification. Our study showed a high concordance of the obtained results, suggesting a good accuracy of the different bioinformatics tools used in the respective centres. In conclusion, inter-centre data exchange is a worthwhile approach for testing and validating the performance of software platforms and the accuracy of results, and is particularly important where the proteomics data contribute to a clinical diagnosis.

Learning temporal attention in dynamic graphs with bilinear interactions.

Reasoning about graphs evolving over time is a challenging concept in many domains, such as bioinformatics, physics, and social networks. We consider a common case in which edges can be short term interactions (e.g., messaging) or long term structural connections (e.g., friendship). In practice, long term edges are often specified by humans. Human-specified edges can be both expensive to produce and suboptimal for the downstream task. To alleviate these issues, we propose a model based on temporal point processes and variational autoencoders that learns to infer temporal attention between nodes by observing node communication. As temporal attention drives between-node feature propagation, using the dynamics of node interactions to learn this key component provides more flexibility while simultaneously avoiding issues associated with human-specified edges. We also propose a bilinear transformation layer for pairs of node features instead of concatenation, typically used in prior work, and demonstrate its superior performance in all cases. In experiments on two datasets in the dynamic link prediction task, our model often outperforms the baseline model that requires a human-specified graph. Moreover, our learned attention is semantically interpretable and infers connections similar to actual graphs.

Artificial cognition: How experimental psychology can help generate explainable artificial intelligence.

Artificial intelligence powered by deep neural networks has reached a level of complexity where it can be difficult or impossible to express how a model makes its decisions. This black-box problem is especially concerning when the model makes decisions with consequences for human well-being. In response, an emerging field called explainable artificial intelligence (XAI) aims to increase the interpretability, fairness, and transparency of machine learning. In this paper, we describe how cognitive psychologists can make contributions to XAI. The human mind is also a black box, and cognitive psychologists have over 150 years of experience modeling it through experimentation. We ought to translate the methods and rigor of cognitive psychology to the study of artificial black boxes in the service of explainability. We provide a review of XAI for psychologists, arguing that current methods possess a blind spot that can be complemented by the experimental cognitive tradition. We also provide a framework for research in XAI, highlight exemplary cases of experimentation within XAI inspired by psychological science, and provide a tutorial on experimenting with machines. We end by noting the advantages of an experimental approach and invite other psychologists to conduct research in this exciting new field.

Comparative study of the stabilities of synthetic in vitro and natural ex vivo transthyretin amyloid fibrils.

Systemic amyloidosis caused by extracellular deposition of insoluble fibrils derived from the pathological aggregation of circulating proteins, such as transthyretin, is a severe and usually fatal condition. Elucidation of the molecular pathogenic mechanism of the disease and discovery of effective therapies still represents a challenging medical issue. The in vitro preparation of amyloid fibrils that exhibit structural and biochemical properties closely similar to those of natural fibrils is central to improving our understanding of the biophysical basis of amyloid formation in vivo and may offer an important tool for drug discovery. Here, we compared the morphology and thermodynamic stability of natural transthyretin fibrils with those of fibrils generated in vitro either using the common acidification procedure or primed by limited selective cleavage by plasmin. The free energies for fibril formation were -12.36, -8.10, and -10.61 kcal mol-1, respectively. The fibrils generated via plasmin cleavage were more stable than those prepared at low pH and were thermodynamically and morphologically similar to natural fibrils extracted from human amyloidotic tissue. Determination of thermodynamic stability is an important tool that is complementary to other methods of structural comparison between ex vivo fibrils and fibrils generated in vitro Our finding that fibrils created via an in vitro amyloidogenic pathway are structurally similar to ex vivo human amyloid fibrils does not necessarily establish that the fibrillogenic pathway is the same for both, but it narrows the current knowledge gap between in vitro models and in vivo pathophysiology.

Three critical factors affecting automated image species recognition performance for camera traps.

Ecological camera traps are increasingly used by wildlife biologists to unobtrusively monitor an ecosystems animal population. However, manual inspection of the images produced is expensive, laborious, and time-consuming. The success of deep learning systems using camera trap images has been previously explored in preliminary stages. These studies, however, are lacking in their practicality. They are primarily focused on extremely large datasets, often millions of images, and there is little to no focus on performance when tasked with species identification in new locations not seen during training. Our goal was to test the capabilities of deep learning systems trained on camera trap images using modestly sized training data, compare performance when considering unseen background locations, and quantify the gradient of lower bound performance to provide a guideline of data requirements in correspondence to performance expectations. We use a dataset provided by Parks Canada containing 47,279 images collected from 36 unique geographic locations across multiple environments. Images represent 55 animal species and human activity with high-class imbalance. We trained, tested, and compared the capabilities of six deep learning computer vision networks using transfer learning and image augmentation: DenseNet201, Inception-ResNet-V3, InceptionV3, NASNetMobile, MobileNetV2, and Xception. We compare overall performance on "trained" locations where DenseNet201 performed best with 95.6% top-1 accuracy showing promise for deep learning methods for smaller scale research efforts. Using trained locations, classifications with <500 images had low and highly variable recall of 0.750 ± 0.329, while classifications with over 1,000 images had a high and stable recall of 0.971 ± 0.0137. Models tasked with classifying species from untrained locations were less accurate, with DenseNet201 performing best with 68.7% top-1 accuracy. Finally, we provide an open repository where ecologists can insert their image data to train and test custom species detection models for their desired ecological domain.

Diagnostic amyloid proteomics: experience of the UK National Amyloidosis Centre.

Systemic amyloidosis is a serious disease which is caused when normal circulating proteins misfold and aggregate extracellularly as insoluble fibrillary deposits throughout the body. This commonly results in cardiac, renal and neurological damage. The tissue target, progression and outcome of the disease depends on the type of protein forming the fibril deposit, and its correct identification is central to determining therapy. Proteomics is now used routinely in our centre to type amyloid; over the past 7 years we have examined over 2000 clinical samples. Proteomics results are linked directly to our patient database using a simple algorithm to automatically highlight the most likely amyloidogenic protein. Whilst the approach has proved very successful, we have encountered a number of challenges, including poor sample recovery, limited enzymatic digestion, the presence of multiple amyloidogenic proteins and the identification of pathogenic variants. Our proteomics procedures and approaches to resolving difficult issues are outlined.