In Silico Peptide Design: Methods, Resources, and Role of AI.

Autistic spectrum disorder (ASD) is a neurodevelopment condition normally linked with substantial healthcare costs and time-consuming assessments where early detection of ASD traits can help limit the development of the condition. The existing conventional ASD screening methods contain a large number of items and are based on domain expert rules which may be criticized of being lengthy and subjective. More importantly, these methods use basic scoring functions to pinpoint to autistic traits rather intelligently learning patterns from cases and controls which can be more accurate and efficient. One promising solution to deal with the above issues and speed up ASD assessment referrals is to develop intelligent artificial intelligence screening methods that not only provide accurate pre-diagnostic classifications but also improve the efficiency and accessibility of the screening process. This paper proposes a new autism screening system that replaces the conventional scoring functions in classic screening methods with deep learning algorithms. The system is composed of a mobile application that provides the user interface capturing questionnaire data; an intelligent ASD detection web service that interfaces with a Convolutional Neural Network (CNN) trained with historical ASD cases; and a database that enables the CNN to learn new knowledge from future users of the system. The CNN classification method was evaluated against a large autism dataset consisting of adult, adolescent, child, and toddler cases and controls. The results obtained from the CNN were compared with other intelligent algorithms in which superior performance was achieved by the CNN. Particularly, the proposed CNN-based ASD classification system revealed higher accuracy, sensitivity, and specificity when compared with conventional screening methods. This indeed will be of high benefit for busy medical clinics and diagnosticians and could possibly be a new direction to change the way ASD diagnosis process is conducted in the future.

Sirolimus is a recently marketed immunosuppressant that, in common with cyclosporine and tacrolimus, exhibits a low average oral bioavailability (approximately 20%). Likewise, sirolimus is a substrate for the major drug-metabolizing enzyme cytochrome P450 3A4 (CYP3A4) and the efflux transporter P-glycoprotein (P-gp), both of which are expressed in close proximity in epithelial cells lining the small intestine. Using CYP3A4-expressing Caco-2 cell monolayers, we examined the interplay between metabolism and transport on the intestinal first-pass extraction of sirolimus. Modified Caco-2 cells metabolized [14C]sirolimus to the same CYP3A4-mediated metabolites as human small intestinal and liver microsomes. [14C]Sirolimus also degraded to the known ring-opened product, seco-sirolimus. A ring-opened dihydro species (M2) was, surprisingly, the major product detected in cells at all sirolimus concentrations examined (2-100 micromol/L) and in incubations with human liver and intestinal homogenates but not in corresponding microsomes. M2 formation was NADPH-dependent but unaffected by prototypical CYP3A4 inhibitors. Although M2 was formed from purified seco-sirolimus (20 micromol/L) in the homogenates, it was not detected in cells when seco-sirolimus was added to the apical compartment because seco-sirolimus was essentially impermeable to the apical membrane. Sirolimus, seco-sirolimus (basolaterally dosed), and M2 were all secreted across the apical membrane, and secretion of each was inhibited by the P-gp inhibitor LY335979 (zosuquidar trihydrochloride). Along with CYP3A4-mediated metabolism and P-gp-mediated efflux, a novel elimination pathway was identified that may also contribute to the first-pass extraction, and hence low oral bioavailability, of sirolimus. This new insight into the intestinal elimination of sirolimus, which was not identified using traditional drug metabolism/transport screening methods, may represent another source for the limited absorption of sirolimus.

BackgroundHarvard Pilgrim Health Care expanded coverage for non-invasive prenatal testing (NIPT) to include all pregnant, single-gestation women aged < 35 years, through a performance-based risk-sharing (PBRS) agreement with Illumina to offset costs from coverage expansion. NIPT analyzes cell-free DNA fragments from a maternal blood sample to screen for fetal aneuploidies and is considered a more accurate screening method than conventional serum biochemical screening and nuchal translucency ultrasound-based approaches.ObjectiveThis study assessed the impact of NIPT coverage expansion on prenatal screening strategies and payer expenditures.MethodsThis was a real-world comparison of utilization and expenditures of prenatal screening and diagnostic testing in pregnant women aged < 35 years pre- (1 March 2016–28 February 2018) and post- (1 March 2018–30 September 2019) coverage expansion. Incidence rate ratios (IRRs) with 95% confidence intervals (CIs) were estimated to compare changes in utilization of conventional and NIPT-based prenatal screening methods. Change in per member per month (PMPM) expenditures in $US year 2020 were assessed post-coverage expansion using a budget impact model.ResultsA total of 5041 and 4109 distinct pregnancies were identified in pre- and post-coverage expansion periods, respectively. Mean ± standard deviation maternal age was consistent between pre- and post-coverage expansion periods (30.35 ± 3.35 and 30.33 ± 3.28, respectively). Screening orders for conventional methods decreased, with an adjusted IRR in the post-expansion period of 0.87 (95% CI 0.85–0.90) times the rate in the pre-expansion period; orders for NIPT increased, with an adjusted IRR in the post-expansion period of 1.41 (95% CI 1.32–1.51) times the rate in the pre-expansion period. Invasive diagnostic testing was low at baseline (1.0%) and did not change post-coverage expansion. The change in PMPM is estimated at $US0.026 post-coverage expansion.ConclusionThe PBRS agreement to expand NIPT coverage for women aged < 35 years was associated with an increase in NIPT utilization, decreases in conventional screening methods, and a modest increase in PMPM expenditures.Supplementary InformationThe online version contains supplementary material available at 10.1007/s41669-021-00261-y.

Artificial intelligence (AI) in the form of deep learning has promise for drug discovery and chemical biology, for example, to predict protein structure and molecular bioactivity, plan organic synthesis, and design molecules de novo. While most of the deep learning efforts in drug discovery have focused on ligand-based approaches, structure-based drug discovery has the potential to tackle unsolved challenges, such as affinity prediction for unexplored protein targets, binding-mechanism elucidation, and the rationalization of related chemical kinetic properties. Advances in deep-learning methodologies and the availability of accurate predictions for protein tertiary structure advocate for a renaissance in structure-based approaches for drug discovery guided by AI. This review summarizes the most prominent algorithmic concepts in structure-based deep learning for drug discovery, and forecasts opportunities, applications, and challenges ahead.

Peptides based on the amino acid sequences found at protein-protein interaction sites make excellent leads for antagonist development. A statistical picture of amino acids involved in protein-protein interactions indicates that proteins recognize and interact with one another through the restricted set of specialized interface amino acid residues, Pro, Ile, Tyr, Trp, Asp and Arg. These amino acids represent residues from each of the three classes of amino acids, hydrophobic, aromatic and charged, with one anionic and one cationic residue at neutral pH. The use of peptides as drug leads has been successfully used to search for antagonists of cell-surface receptors. Peptide, peptidomimetic, and non-peptide organic inhibitors of a class of cell surface receptors, the integrins, currently serve as therapeutic and diagnostic imaging agents. In this review, we discuss the structural features of protein-protein interactions as well as the design of peptides, peptidomimetics, and small organic molecules for the inhibition of protein-protein interactions. Information gained from studying inhibitors of integrin functions is now being applied to the design and testing of inhibitors of other protein-protein interactions. Most drug development progress in the past several decades has been made using the enzyme binding-pocket model of drug targets. Small molecules are designed to fit into the substrate-binding pockets of proteins based on a lock-and-key, induced-fit, or conformational ensemble model of the protein binding site. Traditionally, enzymes have been used as therapeutic drug targets because it was easier to develop rapid, sensitive screening assays, and to find low molecular weight inhibitors that blocked the active site. However, for proteins which interact with other proteins, rather than with small substrate molecules, the lack of binding pockets means that this approach will not generally succeed. There exist many diseases in which the inhibition of protein-protein interactions would provide therapeutic benefit, but there are no general methods available to address such problems. The focus of the first part of this review is to discuss the features of protein-protein interactions which may serve as general guidelines for the development and design of inhibitors for protein-protein interactions. In the second part we focus on the design of peptides (lead compounds) and their conversion into peptidomimetics or small organic molecules for the inhibition of protein-protein interactions. We draw examples from the important and emerging area of integrin-based cell adhesion and show how the principles of protein-protein interactions are followed in the discovery, optimization and usage of specific protein interface peptides as drug leads.

Antimicrobial peptides (AMPs) face stability and toxicity challenges in clinical use. Stapled modification enhances their stability and effectiveness, but its application in peptide design is rarely reported. This study built ten prediction models for stapled AMPs using deep and machine learning, tested their accuracy with an independent data set and wet lab experiments, and characterised stapled loop structures using structural, sequence and amino acid descriptors. AlphaFold improved stapled peptide structure prediction. The support vector machine model performed best, while two deep learning models achieved the highest accuracy of 1.0 on an external test set. Designed cysteine- and lysine-stapled peptides inhibited various bacteria with low concentrations and showed good serum stability and low haemolytic activity. This study highlights the potential of the deep learning method in peptide modification and design.

Urinary tract infections (UTIs) are among the most common bacterial infections in men and urine culture is gold standard for diagnosis. Considering the high prevalence of culture-negative specimens, any method that identifies such specimens is of interest. The aim was to evaluate a new screening concept for flow cytometry analysis (FCA). The outcomes were evaluated against urine culture, uropathogen species and three conventional screening methods. A prospective, consecutive study examined 1,312 urine specimens, collected during January and February 2012. The specimens were analyzed using the Sysmex UF1000i FCA. Based on the FCA data culture negative specimens were identified in a new model by use of linear discriminant analysis (FCA-LDA). In total 1,312 patients were included. In- and outpatients represented 19.6% and 79.4%, respectively; 68.3% of the specimens originated from women. Of the 610 culture-positive specimens, Escherichia coli represented 64%, enterococci 8% and Klebsiella spp. 7%. Screening with FCA-LDA at 95% sensitivity identified 42% (552/1312) as culture negative specimens when UTI was defined according to European guidelines. The proposed screening method was either superior or similar in comparison to the three conventional screening methods. In conclusion, the proposed/suggested and new FCA-LDA screening method was superior or similar to three conventional screening methods. We recommend the proposed screening method to be used in clinic to exclude culture negative specimens, to reduce workload, costs and the turnaround time. In addition, the FCA data may add information that enhance handling and support diagnosis of patients with suspected UTI pending urine culture.

This study addresses the urgent need for an AI model to predict Anaplastic Lymphoma Kinase (ALK) inhibitors for Non-Small Cell Lung Cancer treatment, targeting the ALK-positive mutation. With only five Food and Drug Administration approved ALK inhibitors currently available, effective drugs remain in demand. Leveraging machine learning (ML) and deep learning (DL), our research accelerates the precise screening of novel ALK inhibitors using both ligand-based and structure-based approaches. In ligand-based approach, an ensemble voting model comprising three base learners to classify potential ALK inhibitors, achieving promising retrospective validation results. Notably, the ML-based XGBoost algorithm exhibited compelling results with external validation (EV)-f1 score of 0.921, EV-Average Precision (AP) of 0.961, cross-validation (CV)-f1 score of [Formula: see text] and CV-AP of [Formula: see text]. Besides, the DL-based Artificial Neural Network (ANN) model demonstrated comparative performance with EV-f1 score of 0.930, EV-AP of 0.955, CV-f1 score of [Formula: see text] and CV-AP of [Formula: see text]. For structure-based approach, an XGBoost consensus docking model utilized scores from three molecular docking programs (GNINA 1.0, Vina-GPU 2.0, and AutoDock-GPU) as features. Combining these two approaches, we virtually screened 120,571 compounds, identifying three promising ALK inhibitors, CHEMBL1689515, CHEMBL2380351, and CHEMBL102714, that bind to the protein's pocket and establish hydrophobic contacts in the hinge region through their ketone groups, resembling Alectinib's interaction. Comparative analysis revealed traditional ML models outperformed Graph Neural Networks (GNN), highlighting the critical role of feature engineering and dataset size importance. The study recommends further in vitro testing to validate the prospective screening performance of these models. A graphical user interface is available at https://huggingface.co/spaces/thechuongtrinh/ALK_inhibitors_classification .

Peptides with diverse amino acid sequences, structures, and functions are essential players in biological systems. The construction of well-annotated databases not only facilitates effective information management, search, and mining but also lays the foundation for developing and testing new peptide algorithms and machines. The antimicrobial peptide database (APD) is an original construction in terms of both database design and peptide entries. The host defense antimicrobial peptides (AMPs) registered in the APD cover the five kingdoms (bacteria, protists, fungi, plants, and animals) or three domains of life (bacteria, archaea, and eukaryota). This comprehensive database ( http://aps.unmc.edu/AP ) provides useful information on peptide discovery timeline, nomenclature, classification, glossary, calculation tools, and statistics. The APD enables effective search, prediction, and design of peptides with antibacterial, antiviral, antifungal, antiparasitic, insecticidal, spermicidal, anticancer activities, chemotactic, immune modulation, or antioxidative properties. A universal classification scheme is proposed herein to unify innate immunity peptides from a variety of biological sources. As an improvement, the upgraded APD makes predictions based on the database-defined parameter space and provides a list of the sequences most similar to natural AMPs. In addition, the powerful pipeline design of the database search engine laid a solid basis for designing novel antimicrobials to combat resistant superbugs, viruses, fungi, or parasites. This comprehensive AMP database is a useful tool for both research and education.

COVID-19 detection and classification using chest X-ray images is a current hot research topic based on the important application known as medical image analysis. To halt the spread of COVID-19, it is critical to identify the infection as soon as possible. Due to time constraints and the expertise of radiologists, manually diagnosing this infection from chest X-ray images is a difficult and time-consuming process. Artificial intelligence techniques have had a significant impact on medical image analysis and have also introduced several techniques for COVID-19 diagnosis. Deep learning and explainable AI have shown significant popularity among AL techniques for COVID-19 detection and classification. In this work, we propose a deep learning and explainable AI technique for the diagnosis and classification of COVID-19 using chest X-ray images. Initially, a hybrid contrast enhancement technique is proposed and applied to the original images that are later utilized for the training of two modified deep learning models. The deep transfer learning concept is selected for the training of pretrained modified models that are later employed for feature extraction. Features of both deep models are fused using improved canonical correlation analysis that is further optimized using a hybrid algorithm named Whale-Elephant Herding. Through this algorithm, the best features are selected and classified using an extreme learning machine (ELM). Moreover, the modified deep models are utilized for Grad-CAM visualization. The experimental process was conducted on three publicly available datasets and achieved accuracies of 99.1, 98.2, and 96.7%, respectively. Moreover, the ablation study was performed and showed that the proposed accuracy is better than the other methods.

We describe a general strategy for the design and discovery of affinity peptides for a protein from its natural ligands. Our approach is guided by protein-protein interactions in natural systems and focuses on the hetero-trimeric complex of cardiac troponin I (cTnI), C (cTnC) and T (cTnT). A key premise of this work is that cTnC and cTnT, owing to their innate ability to bind cTnI, are potential templates for the design and discovery of cTnI-binding peptides. Relying only on the knowledge of primary sequences of cTnC and cTnT, we designed a library of short overlapping peptides that span the entirety of cTnC and cTnT and tested them for binding to cTnI. We were successful in identifying several peptides that display high affinity (1-100 nM) for cTnI. The specific implication of this work is that mimicking natural protein-protein interactions is an excellent starting point for the discovery and rational design of peptide ligands. The knowledge of secondary or tertiary structures of the proteins involved is not a necessary precondition for this approach. Nevertheless, we show that structural information can be used to validate the results of a fragment-based peptide design, and can be potentially beneficial for refining the lead candidates. Our approach is broadly applicable to any protein with at least one natural binding ligand with known primary sequence. For protein targets with multiple natural ligands, this approach can potentially yield several distinct affinity peptides capable of simultaneously binding the target protein via orthogonal modes or at complementary interfaces.

In the milieu of promptly advancing technology and increasing demand for electronic devices, circuit defect detection has become crucial to warranting product quality. This study tackles the cons of traditional defect detection methods, proposing a mind-boggling approach based on AI deep learning. The study intends to establish and enhance deep learning algorithms for the exact and real-time detection of circuit defects. This research encompasses an in-depth review of existing literature on circuit defect detection and AI deep learning, underlining the existing gaps and pitfalls in the field. The study will primarily deploy convolutional neural networks (CNNs) and recurrent neural networks (RNNs) as the primary tools to process various data modalities. The results highlight that the proposed AI deep learning framework depicts grander performance, unlike in traditional manual inspection. The study sets precedence in AI applications in quality control as it contributes to improved manufacturing efficiency, reduced production costs, and delivery of utmost-quality electronic products to consumers.

Acquired long QT syndrome caused by drugs that block the human ether-a-go-go-related-gene (hERG) K(+) channel causes severe side effects and thus represents a major problem in clinical studies of drug candidates. Therefore, early prediction of hERG K(+) channel affinity of drug candidates is becoming increasingly important in the drug discovery process. Both structure-based and ligand-based approaches have been undertaken to shed more light on the molecular basis of drug-channel interaction. In this article, in silico approaches for prediction of interaction with hERG are reviewed. Special attention is drawn to the in vitro biological testing systems as well as to consensus approaches for improvement of predictive power.

Acoustic emission source localization is the basic function of the application of acoustic emission technology. For complex structures, mathematical analysis positioning algorithms cannot be obtained, and accurate acoustic emission source localization cannot be obtained, which has always been one of the problems in the actual applications of acoustic emission technology. To solve this problem, this paper proposes an acoustic emission source localization method based on deep learning, which can obtain high-precision acoustic emission source positioning without the need for mathematical analysis positioning algorithms. The AI deep learning acoustic emission source localization method adopts the method of meshing, dividing the grid in the measured structure, generating label data at the grid position, using the label data to conduct AI deep learning training to establish a positioning model, and using the trained model to analyze the actual acoustic emission data (non-labeled data) to locate the acoustic emission source. This paper uses a multi-layer perceptron model to train multi-dimensional features (arrival time and amplitude). This method effectively improves the accuracy of acoustic emission source localization. Experimental results show that the positioning accuracy of the designed deep learning model in the test set reaches 99.625%, which is significantly better than the traditional Time Difference of Arrival (TDOA) positioning algorithm. In addition, this paper further verifies the stability and reliability of the model in localization tasks through credibility metrics such as Score, Margin and Entropy. This article provides a new solution for locating acoustic emission sources in complex structures, and lays a theoretical and practical foundation for the future development of non-destructive testing technology.

This systematic literature review examines the transformative impact of deep learning and explainable AI (XAI) on neurological disease diagnosis using MRI. We analyzed 180 studies from prominent databases, including IEEE Xplore, ScienceDirect, Google Scholar, PubMed, and Scopus, focusing on the methodologies, applications, and emerging trends in diagnosing brain tumors, Alzheimer's disease, and Parkinson's disease. Our findings reveal the increasing use of XAI techniques like Grad-CAM, LIME, and SHAP to enhance model transparency and trustworthiness, a crucial step towards clinical adoption. While deep learning models demonstrate promising diagnostic accuracy, challenges persist, including limited datasets, high computational demands, and the need for robust clinical validation. This review highlights the potential of multimodal data integration and the importance of developing computationally efficient and interpretable models. We identify key future directions, emphasizing the need for larger, more diverse datasets, advancements in XAI methodologies, and the development of personalized treatment strategies guided by AI-driven insights. This comprehensive analysis serves as a valuable resource for researchers and clinicians, offering a roadmap for future research and the responsible implementation of AI in neurological disease diagnosis.