High-level Specification Research Articles

Evaluation of Immunochromatographic Test for Rapid Detection of Norovirus, Rotavirus, and Adenovirus in Stool Samples.

Viral enteric infections are illnesses caused by several types of viruses that primarily affect the gastrointestinal system. Globally, three major viruses associated with gastroenteritis, including norovirus (NoV), rotavirus A (RVA), and human adenovirus (HAdV), have been widely recognized. Accordingly, a rapid and sensitive diagnostic tool, such as immunochromatographic (IC) test, for diarrheal virus detection is required and helpful for rapid diagnosis. The IP-Triple I IC test kit for simultaneous triple virus detections of NoV, RVA, and HAdV was evaluated for its efficacy by using stool specimens that were known positive for these viruses by real-time RT-PCR or PCR, which was used as the gold standard method. The results revealed that the IP-Triple I IC test kit exhibited a very high level of specificity (100%) for NoV, RVA, and HAdV detections, while the sensitivity was slightly different for these three viruses. The sensitivity of detection for RVA was 86.7%, whereas those for NoV and HAdV were 70.6% and 76.2%, respectively. This IP-Triple I IC kit could detect common antigens of a wide range of NoV, RVA, and HAdV genotypes (NoV GII.2, GII.4, GII.6, GII.7, GII.10, GII.17, RVA G1P[8], G2P[4], G3P[8], G8P[8], G9P[8], HAdV-C1, -C2, -C5, -A12, -F40, and -F41). Our results revealed that the IP-Triple I IC test kit is highly effective for simultaneous, direct detection of common antigens of several genotypes of NoV, RVA, and HAdV in stool specimens and useful for screening and rapid diagnosis of diarrheal viruses during the seasonal outbreak.

Advancements in Hyperspectral Imaging and Computer-Aided Diagnostic Methods for the Enhanced Detection and Diagnosis of Head and Neck Cancer.

Background/Objectives: Head and neck cancer (HNC), predominantly squamous cell carcinoma (SCC), presents a significant global health burden. Conventional diagnostic approaches often face challenges in terms of achieving early detection and accurate diagnosis. This review examines recent advancements in hyperspectral imaging (HSI), integrated with computer-aided diagnostic (CAD) techniques, to enhance HNC detection and diagnosis. Methods: A systematic review of seven rigorously selected studies was performed. We focused on CAD algorithms, such as convolutional neural networks (CNNs), support vector machines (SVMs), and linear discriminant analysis (LDA). These are applicable to the hyperspectral imaging of HNC tissues. Results: The meta-analysis findings indicate that LDA surpasses other algorithms, achieving an accuracy of 92%, sensitivity of 91%, and specificity of 93%. CNNs exhibit moderate performance, with an accuracy of 82%, sensitivity of 77%, and specificity of 86%. SVMs demonstrate the lowest performance, with an accuracy of 76% and sensitivity of 48%, but maintain a high specificity level at 89%. Additionally, in vivo studies demonstrate superior performance when compared to ex vivo studies, reporting higher accuracy (81%), sensitivity (83%), and specificity (79%). Conclusion: Despite these promising findings, challenges persist, such as HSI's sensitivity to external conditions, the need for high-resolution and high-speed imaging, and the lack of comprehensive spectral databases. Future research should emphasize dimensionality reduction techniques, the integration of multiple machine learning models, and the development of extensive spectral libraries to enhance HSI's clinical utility in HNC diagnostics. This review underscores the transformative potential of HSI and CAD techniques in revolutionizing HNC diagnostics, facilitating more accurate and earlier detection, and improving patient outcomes.

Evaluation of the Seegene Allplex™ RV master assay for one-step simultaneous detection of eight respiratory viruses in nasopharyngeal specimens

BackgroundThe Seegene Allplex™ RV Master (RVM) assay is a one-step multiplex real-time reverse transcription polymerase chain reaction (RT-PCR) system for detecting eight viral respiratory pathogens from nasopharyngeal swab, aspirate, and bronchoalveolar lavage specimens. The eight RVM targets are: severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), Influenza A (Flu A), Influenza B (Flu B), Human respiratory syncytial virus (RSV), adenovirus (AdV), rhinovirus (HRV), parainfluenza virus (PIV), and metapneumovirus (MPV). The assay is based on Seegene’s multiple detection temperature (MuDT) technology and provides cycle threshold (Ct) values for each of its viral targets upon PCR completion. ObjectiveWe aimed to evaluate the diagnostic performance of the RVM assay by calculating sensitivity, specificity, accuracy, Positive Predictive Value (PPV), Negative Predictive Value (NPV), Positive Percent Agreement (PPA), Negative Percent Agreement (NPA), and Overall Percent Agreement (OPA) compared to definite diagnosis and analogous reference assays. Study designDiagnostic sensitivity, specificity, accuracy, PPV, and NPV were calculated by comparing the results of the RVM assay to that of definite diagnosis assays; while PPA, NPA, and OPA were calculated by comparing results of the RVM assay to that of analogous reference products. Definite diagnosis and reference methods comprised whole genome sequencing and PCR genotyping, the Allplex™ SARS-CoV-2/FluA/FluB/RSV and Respiratory Panels 1, 2, and 3 assays (Seegene), and the Xpert® Xpress SARS-CoV-2/FluA/FluB/RSV Plus assay (Cepheid). Reproducibility of the RVM assay using fully-automated and semi-automated nucleic acid (NA) extraction workflows and as performed by independent operators was also assessed. In total, 249 positive respiratory specimens and at least 50 negative specimens for each target tested were used for this evaluation study. ResultsSensitivity, specificity, accuracy, PPV, NPV, PPA, NPA, and OPA ranged from 95.7 % to 100 % for detecting all eight targets tested on the RVM assay. Reproducibility PPA, NPA, and OPA between automated and semi-automated NA extraction workflows were all >97.9 %, while the reproducibility PPA, NPA and OPA between independent operators were all 100 %. ConclusionThese results demonstrate a high level of sensitivity, specificity, accuracy and diagnostic predictive value of the RVM assay and high agreement with comparable reference assays in identifying all eight of its targets. Taken together, our study underscores the diagnostic utility of the RVM assay in detecting eight viral respiratory pathogens.

Verified Lock-Free Session Channels with Linking

Type systems and program logics based on session types provide powerful high-level reasoning principles for message-passing concurrency. Modern versions employ bidirectional session channels that (1) are asynchronous so that send operations do not block, (2) have buffers in both directions so that both parties can send messages in parallel, and (3) feature a link operation (also called forward) to concisely write programs in process style. These features complicate a low-level lock-free implementation of channels and therefore increase the gap between the meta theory of prior work—which is verified w.r.t. a high-level semantics of channels (e.g., π-calculus)—and the code that runs on an actual computer. We address this problem by verifying a low-level lock-free implementation of session channels w.r.t. a high-level specification based on session types. We carry out our verification in a layered manner by employing the Iris framework for concurrent separation logic. We start with an abstract specification of (unidirectional) queues—of which we provide a linked-list and array-segment based implementation—and gradually build up to session channels with all of the aforementioned features. To make a layered verification possible we develop two logical abstractions—queues with ghost linking and pairing invariants—to reason about the atomicity and changing endpoints due to linking, respectively. All our results are mechanized in the Coq proof assistant.

TRNA-derived fragment tRF-30 propels diabetes-induced retinal microvascular complications by regulating STAT3 signaling.

Transfer RNA-derived fragments (tRFs) represent a novel class of non-coding RNA transcripts that possess specific biological functions. However, the involvement of tRFs in retinal microvascular diseases remains poorly understood. In this study, we aimed to reveal whether modulation of tRF-30 expression could attenuate pathological retinal neovascular diseases. Our findings demonstrate a significant upregulation of tRF-30 expression levels in both in vivo models of diabetic retinopathy (DR) and in vitro endothelial sprouting models. Conversely, inhibition of tRF-30 expression suppressed the formation of abnormal neovascularization in the retina in vivo, while reducing the proliferation and migration activity of retinal vascular endothelial cells in vitro. We also found that tRF-30 modulates retinal neovascularization through the tRF-30/TRIB3/signal transducer and activated transcription 3 signaling pathway. Furthermore, we validated a significant upregulation of tRF-30 expression levels in the vitreous humor of DR patients, with high levels of both validity and specificity in diagnostic testing. Collectively, our findings highlight a pro-angiogenic role for tRF-30 in DR. Intervening in the tRF-30 signaling pathway may represent a promising prevention and treatment strategy for retinal angiogenesis.

Artificial Flowers in the Credit Records of an Eighteenth-Century French Fashion Merchant

Abstract In recent decades, historians have acknowledged the role that women played in shaping and disseminating scientific knowledge during the Enlightenment. Current scholarship also suggests that fashion was a means through which haptic, economic, and practical knowledge was shared among women. This article focuses on one particular fashion accessory – the artificial flower – to explore its contribution to our understanding of women’s knowledge of botany in eighteenth-century France. An analysis of the receipts preserved in the credit records of France’s most famous fashion merchant, Marie-Jeanne (Rose) Bertin (1747–1813), demonstrates high levels of specificity in the flowers that women chose to adorn their outfits. Seventy-five different types of flowers are mentioned using their vernacular names, suggesting that knowledge about a wide variety of flowers was exchanged between fashion merchants and their clients during conversations about clothing. This article therefore casts the fashion merchant’s shop as a site of botanical knowledge generation and exchange.

Peptide-Coated Polycaprolactone-Benzalkonium Chloride Nanocapsules for Targeted Drug Delivery to the Pancreatic β-Cell.

Targeting current therapies to treat or prevent the loss of pancreatic islet β-cells in Type 1 Diabetes (T1D) may provide improved efficacy and reduce off-target effects. Current efforts to target the β-cell are limited by a lack of β-cell-specific targets and the inability to test multiple targeting moieties with the same delivery vehicle. Here, we fabricate a tailorable polycaprolactone nanocapsule (NC) in which multiple different targeting peptides can be interchangeably attached for β-cell-specific delivery. Incorporation of a cationic surfactant in the NC shell allows for the attachment of Exendin-4 and an antibody for ectonucleoside triphosphate diphosphohydrolase 3 (ENTPD3) for β-cell-specific targeting. The average NC size ranges from 250 to 300 nm with a polydispersity index under 0.2. The NCs are nontoxic, stable in media culture, and can be lyophilized and reconstituted. NCs coated with a targeting peptide were taken up by human cadaveric islet β-cells and human stem cell-derived β-like cells (sBC) in vitro with a high level of specificity. Furthermore, NCs successfully delivered both hydrophobic and hydrophilic cargo to human β-cells. Additionally, Exendin-4-coated NCs were stable and targeted the mouse pancreatic islet β-cell in vivo. Overall, our tailorable NCs have the potential to improve cell-targeted drug delivery and can be utilized as a screening platform to test the efficacy of cell-targeting peptides.

Brevetoxin Aptamer Selection and Biolayer Interferometry Biosensor Application.

Brevetoxins (PbTxs) are very potent marine neurotoxins that can cause an illness clinically described as neurologic shellfish poisoning (NSP). These toxins are cyclic polyether in chemistry and have increased their geographical distribution in the past 2 decades. However, the ethical problems as well as technical difficulties associated with currently employed analysis methods for marine toxins have spurred the quest for suitable alternatives to be applied in a regulatory monitoring regime. In this work, we reported the first instance of concurrent aptamer selection of Brevetoxin-1 (PbTx-1) and Brevetoxin-2 (PbTx-2) and constructed a biolayer interferometry (BLI) biosensor utilizing PbTx-1 aptamer as a specific recognition element. Through an in vitro selection process, we have, for the first time, successfully selected DNA aptamers with high affinity and specificity to PbTx-1 and PbTx-2 from a vast pool of random sequences. Among the selected aptamers, aptamer A5 exhibited the strongest binding affinity to PbTx-1, with an equilibrium dissociation constant (KD) of 2.56 μM. Subsequently, we optimized aptamer A5 by truncation to obtain the core sequence (A5-S3). Further refinement was achieved through mutations based on the predictions of a QGRS mapper, resulting in aptamer A5-S3G, which showed a significant increase in the KD value by approximately 100-fold. Utilizing aptamer A5-S3G, we fabricated a label-free, real-time optical BLI aptasensor for the detection of PbTx-1. This aptasensor displayed a broad detection range from 100 nM to 4000 nM PbTx-1, with a linear range between 100 nM and 2000 nM, and a limit of detection (LOD) as low as 4.5 nM. Importantly, the aptasensor showed no cross-reactivity to PbTx-2 or other marine toxins, indicating a high level of specificity for PbTx-1. Moreover, the aptasensor exhibited excellent reproducibility and stability when applied for the detection of PbTx-1 in spiked shellfish samples. We strongly believe that this innovative aptasensor offers a promising alternative to traditional immunological methods for the specific and reliable detection of PbTx-1.

Figurativeness of the Japanese flag

Abstract The analysis offered in this article applies the multilevel approach to conceptual metaphor to relief posters, created and distributed widely after the massive Japan 2011 earthquake. The aim of the analysis is to show in what way the multilevel approach to metaphor can be used to illustrate how the figurative meaning of relief posters emerges from the interactions between different levels of schematicity: from basic image schemas and domains, through contextually embedded semantic frames, finishing with metaphorical scenarios at the highest level of specificity and complexity. The present analysis also shows the potential applicability of image schemas as the guiding analytical element, serving as the departure point for the unfolding metaphorical and metonymical interpretation of a multimodal message. Accordingly, it has been shown how the image schemas of part-whole, full-empty, process and force aid the analysis of metaphoricity and how the central metonymic element of the national flag co-constructs the figurative reading of the posters.

Colorimetric detection of Cu(II) in aqueous solutions using a novel chemosensor: 2-((2-chloroquinolin-3yl)methylidene]amino)phenol

ABSTRACT A spectrophotometric technique based on a selective colorimetric approach is presented in this study for the determination of Cu(II) ions using 2-([(2-chloroquinolin-3-yl)methylidene]amino)phenol (BHMOS) as a detection agent. NMR and FT-IR spectroscopy were used to characterize the compound thoroughly, ensuring accuracy in terms of its molecular structure. As a chemosensor in aqueous solutions, BHMOS has demonstrated a very high level of sensitivity and specificity, even in the presence of other cations. During the interaction between Cu(II) ions and BHMOS, a noticeable colour change was observed to pale green, which is a visual indicator of Cu(II) ions. The method demonstrated linearity for Cu(II) concentrations ranging from 5 to 60 µM, with a detection limit of 1.50 µM. The binding constant for the Cu(II)-BHMOS complex was determined to be 1.14 × 106 M−1, reflecting the strength and stability of the interaction. This method proved to be effective for detecting Cu(II) ions in both natural and simulated water samples, showing its potential utility in environmental monitoring applications.

Enhanced Two-Step LAMP-CRISPR Assay with an Engineered Zst Polymerase for Contamination-Free and Ultrasensitive DNA Detection.

Current loop-mediated isothermal amplification (LAMP)-coupled clustered regularly interspaced short palindromic repeats (LAMP-CRISPR) biosensing in two-step or one-step formats has been applied to next-generation accurate molecular diagnosis. However, two-step LAMP-CRISPR assays intrinsically confront aerosol contamination, while one-step assays possess a compromised detection performance. To this end, we propose an enhanced two-step LAMP-CRISPR assay (ETL-CRISPR) with an engineered Zst polymerase to mediate ultrasensitive DNA detection and thoroughly eliminate aerosol contamination. Instead of supplementing any dTTP, the newly engineered Zst polymerase can efficiently polymerize four oligonucleotides (dATP, dCTP, dGTP, and dUTP), thereby enabling contamination-free and ultrasensitive ETL-CRISPR assay. By targeting the L1 gene of human papillomaviruses (HPV) 16 and the E7 gene of HPV18, our ETL-CRISPR assay achieves high specificity and single-copy level sensitivity within 1 h. Furthermore, we validated the assay by using 85 HPV clinical swab samples with an accuracy of 98.8%, which is comparable to the real-time quantitative polymerase chain reaction. Therefore, ETL-CRISPR provides a straightforward strategy for the contamination-free and ultrasensitive point-of-care diagnosis of clinical pathogens.

Software Synthesis From High-Level Specification for Swarm Robotic Applications

Software Synthesis From High-Level Specification for Swarm Robotic Applications

Integrating untargeted and targeted LC-MS-based metabolomics to identify the serum metabolite biomarkers for tuberculosis.

Given the limitations of untargeted metabolomics in precise metabolite quantification, our current research employed a novel approach by integrating untargeted and targeted metabolomics utilizing ultra-high-performance liquid chromatography quadrupole time-of-flight tandem mass spectrometry (UHPLC-QTOF-MS/MS) to analyze the metabolic profile and potential biomarkers for tuberculosis (TB). A cohort of 36 TB patients and 36 healthy controls (HC) was enlisted to obtain serum samples. Multivariate pattern recognition and univariate statistical analysis were employed to screen and elucidate the differential metabolites, whereas dot plots and receiver operating characteristic (ROC) curves were established for the identification of potential biomarkers of TB. The results indicated a distinct differentiation between the two groups, identifying 99 metabolites associated with five primary metabolic pathways in relation to TB. Of these, 19 metabolites exhibited high levels of sensitivity and specificity, as evidenced by the area under curve values approaching 1. Following targeted quantitative analysis, three potential metabolites, namely, L-asparagine, L-glutamic acid, and arachidonic acid, were demonstrated excellent discriminatory ability as evidenced by the results of the ROC curve, dot plots, and random forest model. Particularly noteworthy was the enhanced diagnostic efficacy of the combination of these three metabolites compared to singular biomarkers, suggesting their potential utility as serum biomarkers for TB diagnosis.

Novel Discrete and Imprinted Fluoride-Selective Sensors: Bridging the Gap from DMSO to Aqueous Samples.

Fluoride in drinking water has beneficial or harmful health effects depending on its concentration. This highlights the need for new low-cost and portable sensors capable of in situ monitoring of F- ions. Unfortunately, achieving high levels of water compatibility and fluoride specificity remains a challenge. Here, four new urea-based discrete sensors are prepared and characterized. The sensors containing anthracenyl- (5) and 9H-fluorenyl- (7) signaling units exhibit intense luminescent emissions in dimethyl sulfoxide, the former being particularly sensitive and selective to fluoride. In water, 5 displays a superior sensitivity (871M-1) and a detection limit (8µm) below international guidelines, albeit with cross-sensitivity to H2PO4‾. To enhance the performance, 5 and 7 are embedded into a fluoride-imprinted polymeric matrix to give solid-state sensors (5P and 7P, respectively). 5P shows good sensitivity (360M-1) and specificity in water. Besides, it has a low detection limit (35µm) and a response linear range (118-6300µm) encompassing the limit established by the Environmental Protection Agency (211µm). Furthermore, 5P also displays good reusability and adequate recovery values in real-sample testing (102±2%), constituting the first example of a low-cost anion-imprinted polymeric probe tailored for the selective sensing of fluoride in aqueous samples.

Artificial Intelligence for the Automatic Diagnosis of Gastritis: A Systematic Review.

Background and Objective: Gastritis represents one of the most prevalent gastrointestinal diseases and has a multifactorial etiology, many forms of manifestation, and various symptoms. Diagnosis of gastritis is made based on clinical, endoscopic, and histological criteria, and although it is a thorough process, many cases are misdiagnosed or overlooked. This systematic review aims to provide an extensive overview of current artificial intelligence (AI) applications in gastritis diagnosis and evaluate the precision of these systems. This evaluation could highlight the role of AI as a helpful and useful tool in facilitating timely and accurate diagnoses, which in turn could improve patient outcomes. Methods: We have conducted an extensive and comprehensive literature search of PubMed, Scopus, and Web of Science, including studies published until July 2024. Results: Despite variations in study design, participant numbers and characteristics, and outcome measures, our observations suggest that implementing an AI automatic diagnostic tool into clinical practice is currently feasible, with the current systems achieving high levels of accuracy, sensitivity, and specificity. Our findings indicate that AI outperformed human experts in most studies, with multiple studies exhibiting an accuracy of over 90% for AI compared to human experts. These results highlight the significant potential of AI to enhance diagnostic accuracy and efficiency in gastroenterology. Conclusions: AI-based technologies can now automatically diagnose using images provided by gastroscopy, digital pathology, and radiology imaging. Deep learning models exhibited high levels of accuracy, sensitivity, and specificity while assessing the diagnosis, staging, and risk of neoplasia for different types of gastritis, results that are superior to those of human experts in most studies.

Sequence-specific nanoparticle barcode strategy for multiplex human enterovirus typing

Human enteroviruses (HEV) can cause a range of diseases from mild to potentially life-threatening. Identification and genotyping of HEV are crucial for disease management. Existing typing methods, however, have inherent limitations. Developing alternative methods to detect HEV with more virus types, high accuracy, and sensitivity in an accessible manner presents a technological and analytical challenge. Here, a sequence-specific nanoparticle barcode (SSNB) method is presented for simultaneous detection of 10 HEV types. This method significantly increases sensitivity, enhancing detection by 10-106 times over the traditional multiplex hybrid genotyping (MHG) method, by resolving cross-interference between the multiple primer sets. Furthermore, the SSNB method demonstrates a 100% specificity in accurately distinguishing between 10 different HEV types and other prevalent clinical viruses. In an analysis of 70 clinical throat swab samples, the SSNB method shows slightly higher detection rate for positive samples (50%) compared to the RT-PCR method (48.6%). Additionally, further assessment of the typing accuracy for samples identified as positive by SSNB using sequencing method reveals a concordance rate of 100%. The combined high sensitivity and specificity level of the methodology, together with the capability for multiple type analysis and compatibility with clinical workflow, make this approach a promising tool for clinical settings.

The Utilization of Brillouin Microscopy in Corneal Diagnostics: A Systematic Review.

Corneal biomechanical data has been used since 2005 to screen for keratoconus and corneal ectasia by corneal specialists. Older technology uses force applanation techniques over a 3 mm area in the central cornea, making it highly dependent on extraneous variables and unable to calculate the elasticity of the tissue. Brillouin microscopy is a newer method that uses a natural shift in the frequency of light as it passes through a material. This frequency shift can be used to estimate the viscoelasticity of the tissue. The advantage of Brillouin microscopy is that it can create a full three-dimensional (3D) map of the entire cornea without direct contact. A literature search was conducted using the databases PubMed, Google Scholar, and Ovid regarding the applications of Brillouin microscopy in corneal diagnostics. A final total of 16 articles was included describing the various ex vivo and in vivo studies conducted using Brillouin microscopy. Applications of this technology spanned from keratoconus diagnosis to post-corneal refractive surgery evaluation. All studies evaluated corneal biomechanics and other corneal properties through the quantification of Brillouin frequency shifts. Many of the studies found that this diagnostic device is capable of detecting subtle changes in corneal thickness and biomechanics in keratoconic corneas at a high level of specificity and sensitivity. However, limitations of Brillouin microscopy may include the duration of time required for use and fluctuations in accuracy depending on the corneal hydration state. Future technology seems to be geared towarda combination of optical coherence tomography (OCT) and Brillouin microscopy, using OCT as a three-dimensional pupil-tracking modality. Further research and understanding of the technology involved will lead to better care of patients in the field of ophthalmology.

Relation-Algebraic Verification of Disjoint-Set Forests

This paper studies how to use relation algebras, which are useful for high-level specification and verification, for proving the correctness of lower-level array-based implementations of algorithms. We give a simple relation-algebraic semantics of read and write operations on associative arrays. The array operations seamlessly integrate with assignments in computation models supporting while-programs. As a result, relation algebras can be used for verifying programs with associative arrays. We verify the correctness of an array-based implementation of disjoint-set forests using the union-by-rank strategy and find operations with path compression, path splitting and path halving. All results are formally proved in Isabelle/HOL. This paper is an extended version of [1].

Применение машинного обучения к данным нейроанатомии и физиологии в области диагностики СДВГ

<p style="text-align: justify;">Attention Deficit\Hyperactivity Disorder (ADHD) is a common neurodevelopmental disorder characterized by inattention, hyperactivity, and impulsivity. While traditional diagnostic methods rely on clinical interviews, tests and behavioral observations, machine learning methods provide an opportunity to simplify the ADHD diagnostic process and make it more accurate. This review tries to explore the application of machine learning (ML) algorithms to physiological and neuroanatomical data: magnetic resonance imaging (MRI), functional MRI (fMRI), near-infrared spectroscopy (fNIRS), electroencephalography (EEG), magnetoencephalography (MEG), electrocardiogram (ECG), pupil parameters, eye tracking and activity in the field of exploring biomarkers for ADHD diagnosis. Deep learning models and support vector machines (SVM) are considered the most promising approaches for identifying ADHD in both children and adults. However, despite the fact that with the help of machine learning methods researchers are able to achieve high levels of specificity and sensitivity when solving problems of ADHD assessment, their use in clinical practice requires preliminary work to verify the results on large samples, as well as addressing data security and ethical issues.</p>

ToxinPred 3.0: An improved method for predicting the toxicity of peptides

Toxicity emerges as a prominent challenge in the design of therapeutic peptides, causing the failure of numerous peptides during clinical trials. In 2013, our group developed ToxinPred, a computational method that has been extensively adopted by the scientific community for predicting peptide toxicity. In this paper, we propose a refined variant of ToxinPred that showcases improved reliability and accuracy in predicting peptide toxicity. Initially, we utilized a similarity/alignment-based approach employing BLAST to predict toxic peptides, which yielded satisfactory accuracy; however, the method suffered from inadequate coverage. Subsequently, we employed a motif-based approach using MERCI software to uncover specific patterns or motifs that are exclusively observed in toxic peptides. The search for these motifs in peptides allowed us to predict toxic peptides with a high level of specificity with poor sensitivity. To overcome the coverage limitations, we developed alignment-free methods using machine/deep learning techniques to balance sensitivity and specificity of prediction. Deep learning model (ANN - LSTM with fixed sequence length) developed using one-hot encoding achieved a maximum AUROC of 0.93 with MCC of 0.71 on an independent dataset. Machine learning model (extra tree) developed using compositional features of peptides achieved a maximum AUROC of 0.95 with MCC of 0.78. We also developed large language models and achieved maximum AUC of 0.93 using ESM2-t33. Finally, we developed hybrid or ensemble methods combining two or more methods to enhance performance. Our specific hybrid method, which combines a motif-based approach with a machine learning-based model, achieved a maximum AUROC of 0.98 with MCC 0.81 on an independent dataset. In this study, all models were trained and tested on 80 % of data using five-fold cross-validation and evaluated on the remaining 20 % of data called independent dataset. The evaluation of all methods on an independent dataset revealed that the method proposed in this study exhibited better performance than existing methods. To cater to the needs of the scientific community, we have developed a standalone software, pip package and web-based server ToxinPred3 (https://github.com/raghavagps/toxinpred3 and https://webs.iiitd.edu.in/raghava/toxinpred3/).