ToxIBTL: prediction of peptide toxicity based on information bottleneck and transfer learning

PLPTP: A Motif-based Interpretable Deep Learning Framework Based on Protein Language Models for Peptide Toxicity Prediction

Peptide toxicity prediction holds significant importance in drug development and biotechnology, as accurately identifying toxic peptide sequences is crucial for designing safer peptide-based drugs. This study proposes a deep learning-based model for peptide toxicity prediction, integrating Evolutionary Scale Modeling (ESM2), Bidirectional Long Short-Term Memory (BiLSTM), and Deep Neural Network (DNN). The ESM2 model captures evolutionary information from peptide sequences, providing a rich context for the sequences; the BiLSTM network focuses on extracting contextual dependencies, thereby capturing long-range dependencies within the sequence; and the DNN further classifies the extracted features to achieve the final toxicity prediction. To enhance the reliability and transparency of the model, we also conducted motif analysis to identify key patterns in the data, which helps to explain the model's attention mechanism and its classification performance. To address the class imbalance in the dataset, we employed Focal Loss as the loss function, which enhances the model's ability to identify minority class samples by reducing the contribution of easily classified samples. Experimental results demonstrate that the proposed model performs exceptionally well across multiple evaluation metrics, particularly in handling imbalanced data, achieving significant improvements over traditional methods. This result highlights the model's potential to improve the accuracy of peptide toxicity prediction and its valuable role in drug development and biotechnology research. The PLPTP web server is available at https://www.bioai-lab.com/PLPTP.

Identification of bioaccessible and neuroprotective peptides from fermented casein hydrolysate

Fermented dairy products are beneficial to cognitive health. Fermentation-released bioactive peptides have the potential to contribute to the neuroprotective effects of fermented dairy products. However, known neuroprotective peptides are mostly prepared by enzymatic hydrolysis, and physicochemical screening of food-derived functional peptides typically overlooks the interference of biotransport after ingestion. Thus, we aimed to identify neuroprotective peptides from casein fermented by Lactobacillus delbrueckii ssp. bulgaricus to provide more evidence supporting the contribution of fermentation-released peptides. We first screened bioaccessible peptides from fermented casein hydrolysate by simulating digestion, absorption, and blood-brain barrier penetration using INFOGEST standardized protocols, human colon Caco-2 cells, and human brain microvascular endothelial hCMEC/D3 cells sequentially. Next, we identified peptides of each stage by nano-liquid chromatography tandem MS. The intersections were considered bioaccessible peptides. We performed molecular docking against Kelch-like ECH-associated protein 1 (Keap1) to predict potential bioactive peptides and validated the predicted effects in BV2 microglial cells induced by LPS. As a result, we identified 1,971, 663, 276, and 208 casein peptides from the simulated products at each stage, and 63 bioaccessible peptides were identified during fermentation, underwent simulated digestion, and were transported via the simulated intestinal epithelial barrier and blood-brain barrier. Among these peptides, 7 nontoxic small peptides had relatively high predicted affinities for Keap1 and were verified in LPS-treated BV2 cells. We found that Phe-Val-Ala-Pro-Phe-Pro-Glu (FE7) decreased nitric oxide, IL-1β, reactive oxygen species, and lipid peroxidation levels by 69.6%, 103.6%, 119.3%, and 75.3%, respectively, in LPS-treated BV2 cells. In conclusion, FE7 could be a promising neuroprotective peptide in fermented casein hydrolysate by reducing neuroinflammation and oxidative stress. Our approach provides a feasible paradigm for identifying bioaccessible and neuroprotective peptides from dairy products.

An in silico vaccinomics strategy to develop multiepitope vaccine using essential hypothetical protein as a target against Brevundimonas subvibrioides: A combined subtractive proteomics and immunoinformatics approach

The genus Brevundimonas, responsible for a spectrum of diseases, encompasses opportunistic pathogens against diverse infections, with B. subvibrioides possessing the largest genome size. Presence of Brevundimonas genus has been found in clinical samples of patients with urinary tract infections (UTIs) and mastitis, specifically, B. subvibrioides. However, former treatment methods using antibiotics have rendered the bacteria resistant to the inhibitory effects of chloramphenicol, azithromycin, etc. as well as alternative chemical treatments. Reckoning with the present scenario, vaccination emerges as the safest and most effective treatment strategy to address the global issue of evolving microbial threats. The growing concern of antibiotic resistance necessitates a shift towards vaccination as the primary treatment strategy. For this study, a total of 15 essential hypothetical proteins (EsHPs) were retrieved from the database of essential genes (DEG) for further multi-server functional annotation, physicochemical characterization, and non-homology analysis. The target antigen for our peptide-based vaccine construct was localized in the extracellular space, exhibited virulence factor and non-homology against human host and gut microflora. The final screened candidate was detected to be antigenic, probable non-allergen, soluble, non-toxic with near absence of transmembrane helices. The linear B-cell lymphocyte (LBL), helper T lymphocyte (HTL), and cytotoxic T lymphocyte (CTL) epitopes present on this protein were predicted and further docked with their respective major histocompatibility complex (MHC) molecules to verify their affinities for favourable antigen presentation. The final vaccine construct was built with human beta-defensin 3 (HBD3) as the adjuvant, 2 LBL epitopes, 3 CTL epitopes, 1 HTL epitope, and the required linkers. The designed construct was found to be immune response-inducing and was therefore cloned into a suitable vector to assess the possibility of appropriate expression. The final construct, when docked with Toll-like receptor 4 (TLR4), followed by molecular dynamics (MD) simulation of unbound and bound complexes, demonstrated a strong generation of immunity, whose safety might be experimentally tested in the near future. This was corroborated by stabilizing root-mean-square deviation (RMSD) curves, increasing hydrogen bonds, and decreasing residual mobility.

Immunoinformatics-driven design of a multi-epitope vaccine targeting neonatal rotavirus with focus on outer capsid proteins VP4 and VP7 and non structural proteins NSP2 and NSP5

Rotaviral gastroenteritis remains a major global health concern, particularly for infants and young children under five years old. Prior to the introduction of the WHO-prequalified rotavirus vaccine, rotavirus (RV) was responsible for approximately 800,000 child deaths annually in developing countries. Although vaccination efforts have reduced this number, RV still causes around 200,000 child deaths each year worldwide. The current WHO-prequalified vaccines are live attenuated and offer limited efficacy of 40-60%, with a slight risk of intussusception in young children. To overcome these limitations, we employed immunoinformatics to design a novel multi-epitope vaccine (MEV) targeting rotavirus outer capsid proteins VP4 and VP7, as well as crucial non-structural proteins NSP2 and NSP5. The RV-MEV incorporates 10 epitopes, including 4 CD8 + T-cell, 5 CD4 + T-cell, and 1 B-cell epitope, all of which are antigenic, non-allergenic, and non-toxic. These epitopes also showed potential to induce interferon-γ (IFN-γ). Molecular simulation studies confirmed stable interactions between RV-MEV and human TLR5 and integrin αvβ5 complexes. The RV-MEV was successfully cloned into a pET28a(+) vector during in-silico cloning. Immune simulation studies predict a strong immune response to the RV-MEV. Future in vitro and in vivo studies are necessary to validate the vaccine's effectiveness in providing protection against various rotavirus strains in neonates.

ToxDL 2.0: Protein toxicity prediction using a pretrained language model and graph neural networks

Motivation

Assessing the potential toxicity of proteins is crucial for both therapeutic and agricultural applications. Traditional experimental methods for protein toxicity evaluation are time-consuming, expensive, and labor-intensive, highlighting the requirement for efficient computational approaches. Recent advancements in language models and deep learning have significantly improved protein toxicity prediction, yet current models often lack the ability to integrate evolutionary and structural information, which is crucial for accurate toxicity assessment of proteins.

Results

In this study, we present ToxDL 2.0, a novel multimodal deep learning model for protein toxicity prediction that integrates both evolutionary and structural information derived from a pretrained language model and AlphaFold2. ToxDL 2.0 consists of three key modules: (1) a Graph Convolutional Network (GCN) module for generating protein graph embeddings based on AlphaFold2-predicted structures, (2) a domain embedding module for capturing protein domain representations, and (3) a dense module that combines these embeddings to predict the toxicity. After constructing a comprehensive toxicity benchmark dataset, we obtained experimental results on both an original non-redundant test set (comprising pre-2022 protein sequences) and an independent non-redundant test set (a holdout set of post-2022 protein sequences), demonstrating that ToxDL 2.0 outperforms existing state-of-the-art methods. Additionally, we utilized Integrated Gradients to discover known toxic motifs associated with protein toxicity. A web server for ToxDL 2.0 is publicly available at www.csbio.sjtu.edu.cn/bioinf/ToxDL2/.

Machine Learning-Enabled Drug-Induced Toxicity Prediction

Unexpected toxicity has become a significant obstacle to drug candidate development, accounting for 30% of drug discovery failures. Traditional toxicity assessment through animal testing is costly and time-consuming. Big data and artificial intelligence (AI), especially machine learning (ML), are robustly contributing to innovation and progress in toxicology research. However, the optimal AI model for different types of toxicity usually varies, making it essential to conduct comparative analyses of AI methods across toxicity domains. The diverse data sources also pose challenges for researchers focusing on specific toxicity studies. In this review, 10 categories of drug-induced toxicity is examined, summarizing the characteristics and applicable ML models, including both predictive and interpretable algorithms, striking a balance between breadth and depth. Key databases and tools used in toxicity prediction are also highlighted, including toxicology, chemical, multi-omics, and benchmark databases, organized by their focus and function to clarify their roles in drug-induced toxicity prediction. Finally, strategies to turn challenges into opportunities are analyzed and discussed. This review may provide researchers with a valuable reference for understanding and utilizing the available resources to bridge prediction and mechanistic insights, and further advance the application of ML in drugs-induced toxicity prediction.

Computational Development of Allosteric Peptide Inhibitors Targeting LIM Kinases as a Novel Therapeutic Intervention

LIM Kinases (LIMKs) have emerged as critical therapeutic targets in cancer research due to their central role in regulating cytoskeletal dynamics and cell motility via cofilin phosphorylation. Allosteric inhibitors, which bind outside the ATP-binding pocket, offer distinct advantages over ATP-competitive inhibitors, such as increased specificity, reduced off-target effects, and the ability to overcome resistance. This study investigates a series of novel tetrapeptides mimicking the binding mode of TH470, an allosteric LIMK inhibitor, using in silico docking and molecular dynamics simulations to identify potential lead compounds with high specificity, binding affinity, and favorable pharmacokinetic properties. Structural analyses revealed critical interactions between TH470 and LIMKs, particularly with conserved residues such as Thr405 (gatekeeper residue), Ile408 (hinge region), and Asp469 (XDFG motif), which are essential for stabilizing inhibitor binding. Molecular dynamics simulations confirmed the stability of TH470-LIMK1 and TH470-LIMK2 complexes, with lower RMS deviations and robust interaction patterns enhancing binding affinity. From the set of tetrapeptides mimicking TH470 binding mode, only YFYW, WPHW, and YWFP for LIMK1, and PYWG, FYWV, and WFVW for LIMK2 demonstrated high binding affinities, non-toxic profiles, and promising anti-cancer, anti-angiogenic, and anti-inflammatory properties. Among the studied peptides, LIMK1-YFYW and LIMK2-WFVW exhibited the most substantial binding affinities, supported by high hydrogen bond occupancy with key residues such as Ile416 and Thr405. The findings highlight the therapeutic potential of allosteric peptide inhibitors targeting LIMK-mediated pathways in cancer progression. The study underscores the importance of specific interactions with conserved LIMK residues, providing a foundation for further developing selective inhibitors to modulate actin dynamics and combat cancer-related processes.

DeepPD: A Deep Learning Method for Predicting Peptide Detectability Based on Multi-feature Representation and Information Bottleneck

Peptide detectability measures the relationship between the protein composition and abundance in the sample and the peptides identified during the analytical procedure. This relationship has significant implications for the fundamental tasks of proteomics. Existing methods primarily rely on a single type of feature representation, which limits their ability to capture the intricate and diverse characteristics of peptides. In response to this limitation, we introduce DeepPD, an innovative deep learning framework incorporating multi-feature representation and the information bottleneck principle (IBP) to predict peptide detectability. DeepPD extracts semantic information from peptides using evolutionary scale modeling 2 (ESM-2) and integrates sequence and evolutionary information to construct the feature space collaboratively. The IBP effectively guides the feature learning process, minimizing redundancy in the feature space. Experimental results across various datasets demonstrate that DeepPD outperforms state-of-the-art methods. Furthermore, we demonstrate that DeepPD exhibits competitive generalization and transfer learning capabilities across diverse datasets and species. In conclusion, DeepPD emerges as the most effective method for predicting peptide detectability, showcasing its potential applicability to other protein sequence prediction tasks.

Modern machine learning methods for protein property prediction

Recent progress and development of artificial intelligence and machine learning (AI/ML) techniques have enabled addressing complex biomolecular problems. AI/ML models learn the underlying distribution of data they are trained on and when exposed to new inputs, they make predictions based on patterns and relationships previously observed in the training set. Further, generative artificial intelligence (GenAI) can be used to accurately generate protein structure or sequence from specific selected properties. This review specifically focuses on the applications of AI/ML in predicting important functional properties of proteins, and the potential prospects of reverse-engineering in depicting the sequence and structure, from available protein-property information.

Dual-Labeled Small Peptides in Cancer Imaging and Fluorescence-Guided Surgery: Progress and Future Perspectives

Dual-labeled compounds that combine radiolabeling and fluorescence labeling represent a significant advancement in precision oncology. Their clinical implementation enhances patient care and outcomes by leveraging the high sensitivity of radioimaging for tumor detection and taking advantage of fluorescence-based optical visualization for surgical guidance. Non-invasive radioimaging facilitates immediate identification of both primary tumors and metastases, while fluorescence imaging assists in decision-making during surgery by offering a spatial distinction between malignant and non-malignant tissue. These advancements hold promise for enhancing patient outcomes and personalization of cancer treatment. The development of dual-labeled molecular probes targeting various cancer biomarkers is crucial in addressing the heterogeneity inherent in cancer pathology and recent studies had already demonstrated the impact of dual-labeled compounds in surgical decision-making (NCT03699332, NCT03407781). This review focuses on the development and application of small dual-labeled peptides in the imaging and treatment of various cancer types. It summarizes the biomarkers targeted to date, tracing their development from initial discovery to the latest advancements in peptidomimetics. Through comprehensive analysis of recent preclinical and clinical studies, the review demonstrates the potential of these dual-labeled peptides to improve tumor detection, localization, and resection. Additionally, it highlights the evolving landscape of dual-modality imaging, emphasizing its critical role in advancing personalized and effective cancer therapy. This synthesis of current research underscores the promise of dual-labeled peptides in enhancing diagnostic accuracy and therapeutic outcomes in oncology.

AI Methods for Antimicrobial Peptides: Progress and Challenges

Antimicrobial peptides (AMPs) are promising candidates to combat multidrug-resistant pathogens. However, the high cost of extensive wet-lab screening has made AI methods for identifying and designing AMPs increasingly important, with machine learning (ML) techniques playing a crucial role. AI approaches have recently revolutionised this field by accelerating the discovery of new peptides with anti-infective activity, particularly in preclinical mouse models. Initially, classical ML approaches dominated the field, but recently there has been a shift towards deep learning (DL) models. Despite significant contributions, existing reviews have not thoroughly explored the potential of large language models (LLMs), graph neural networks (GNNs) and structure-guided AMP discovery and design. This review aims to fill that gap by providing a comprehensive overview of the latest advancements, challenges and opportunities in using AI methods, with a particular emphasis on LLMs, GNNs and structure-guided design. We discuss the limitations of current approaches and highlight the most relevant topics to address in the coming years for AMP discovery and design.

Rore: robust and efficient antioxidant protein classification via a novel dimensionality reduction strategy based on learning of fewer features

In protein identification, researchers increasingly aim to achieve efficient classification using fewer features. While many feature selection methods effectively reduce the number of model features, they often cause information loss caused by merely selecting or discarding features, which limits classifier performance. To address this issue, we present Rore, an algorithm based on a feature-dimensionality reduction strategy. By mapping the original features to a latent space, Rore retains all relevant feature information while using fewer representations of the latent features. This approach significantly preserves the original information and overcomes the information loss problem associated with previous feature selection. Through extensive experimental validation and analysis, Rore demonstrated excellent performance on an antioxidant protein dataset, achieving an accuracy of 95.88% and MCC of 91.78%, using vectors including only 15 features. The Rore algorithm is available online at http://112.124.26.17:8021/Rore .

Peptide nanozymes: An emerging direction for functional enzyme mimics

The abundance of molecules on early Earth likely enabled a wide range of prebiotic chemistry, with peptides playing a key role in the development of early life forms and the evolution of metabolic pathways. Among peptides, those with enzyme-like activities occupy a unique position between peptides and enzymes, combining both structural flexibility and catalytic functionality. However, their full potential remains largely untapped. Further exploration of these enzyme-like peptides at the nanoscale could provide valuable insights into modern nanotechnology, biomedicine, and even the origins of life. Hence, this review introduces the groundbreaking concept of "peptide nanozymes (PepNzymes)", which includes single peptides exhibiting enzyme-like activities, peptide-based nanostructures with enzyme-like activities, and peptide-based nanozymes, thus enabling the investigation of biological phenomena at nanoscale dimensions. Through the rational design of enzyme-like peptides or their assembly with nanostructures and nanozymes, researchers have found or created PepNzymes capable of catalyzing a wide range of reactions. By scrutinizing the interactions between the structures and enzyme-like activities of PepNzymes, we have gained valuable insights into the underlying mechanisms governing enzyme-like activities. Generally, PepNzymes play a crucial role in biological processes by facilitating small-scale enzyme-like reactions, speeding up molecular oxidation-reduction, cleavage, and synthesis reactions, leveraging the functional properties of peptides, and creating a stable microenvironment, among other functions. These discoveries make PepNzymes useful for diagnostics, cellular imaging, antimicrobial therapy, tissue engineering, anti-tumor treatments, and more while pointing out opportunities. Overall, this research provides a significant journey of PepNzymes' potential in various biomedical applications, pushing them towards new advancements.

Unlocking the Mycobacteroides abscessus pan-genome using computational tools: insights into evolutionary dynamics and lifestyle

Mycobacteroides abscessus is a non-tuberculous mycobacteria implicated in causing lung infections. It is difficult to control owing to resistance to antibiotics and disinfectants. This work was aimed at comprehending: the pan-genome architecture, evolutionary dynamics, and functionalities of pan-genome components linked to COGs and KEGG. Around 2802 core genes were present in each strain of the M. abscessus genome. The number of accessory genes ranged from 1615 to 2481. The open pan-genome of M. abscessus was attributed to the accessory genes underlining its adaptability in the host. Phylogenetic analysis revealed cluster-based relationships and highlighted factors shaping variability and adaptive capabilities. Transcription, metabolism, and pathogenic genes were vital for M. abscessus lifestyle. The accessory genes contributed to the diverse metabolic capability. The incidence of a significant portion of secondary metabolite biosynthesis genes provided insights for investigating their biosynthetic gene clusters. Additionally, a high proportion of xenobiotic biodegradation genes highlighted potential metabolic capabilities. In silico screening identified a potential vaccine candidate among hypothetical proteins in COGs. Functional analysis of M. abscessus pan-genome components unveiled factors associated with virulence, pathogenicity, infection establishment, persistence, and resistance. Notable amongst them were: MMPL family transporters, PE-PPE domain-containing proteins, TetR family transcriptional regulators, ABC transporters, Type-I, II, III, VII secretion proteins, DUF domain-containing proteins, cytochrome P450, VapC family toxin, virulence factor Mce family protein, type II toxin-antitoxin system. Overall, these results enhanced understanding of the metabolism, host-pathogen dynamics, pathogenic lifestyle, and adaptations. This will facilitate further investigations for combating infections and designing suitable therapies.

Advances of computational methods enhance the development of multi-epitope vaccines

Vaccine development is one of the most promising fields, and multi-epitope vaccine, which does not need laborious culture processes, is an attractive alternative to classical vaccines with the advantage of safety, and efficiency. The rapid development of algorithms and the accumulation of immune data have facilitated the advancement of computer-aided vaccine design. Here we systemically reviewed the in silico data and algorithms resource, for different steps of computational vaccine design, including immunogen selection, epitope prediction, vaccine construction, optimization, and evaluation. The performance of different available tools on epitope prediction and immunogenicity evaluation was tested and compared on benchmark datasets. Finally, we discuss the future research direction for the construction of a multiepitope vaccine.

Hyb_SEnc: An Antituberculosis Peptide Predictor Based on a Hybrid Feature Vector and Stacked Ensemble Learning

Tuberculosis has plagued mankind since ancient times, and the struggle between humans and tuberculosis continues. Mycobacterium tuberculosis is the leading cause of tuberculosis, infecting nearly one-third of the world's population. The rise of peptide drugs has created a new direction in the treatment of tuberculosis. Therefore, for the treatment of tuberculosis, the prediction of anti-tuberculosis peptides is crucial. This paper proposes an anti-tuberculosis peptide prediction method based on hybrid features and stacked ensemble learning. First, a random forest (RF) and extremely randomized tree (ERT) are selected as first-level learning of stacked ensembles. Then, the five best-performing feature encoding methods are selected to obtain the hybrid feature vector, and then the decision tree and recursive feature elimination (DT-RFE) are used to refine the hybrid feature vector. After selection, the optimal feature subset is used as the input of the stacked ensemble model. At the same time, logistic regression (LR) is used as a stacked ensemble secondary learner to build the final stacked ensemble model Hyb_SEnc. The prediction accuracy of Hyb_SEnc achieved 94.68% and 95.74% on the independent test sets of AntiTb_MD and AntiTb_RD, respectively.

Unveiling Encrypted Antimicrobial Peptides from Cephalopods' Salivary Glands: A Proteolysis-Driven Virtual Approach

Antimicrobial peptides (AMPs) have potential against antimicrobial resistance and serve as templates for novel therapeutic agents. While most AMP databases focus on terrestrial eukaryotes, marine cephalopods represent a promising yet underexplored source. This study reveals the putative reservoir of AMPs encrypted within the proteomes of cephalopod salivary glands via in silico proteolysis. A composite protein database comprising 5,412,039 canonical and noncanonical proteins from salivary apparatus of 14 cephalopod species was subjected to digestion by 5 proteases under three protocols, yielding over 9 million of nonredundant peptides. These peptides were effectively screened by a selection of 8 prediction and sequence comparative tools, including machine learning, deep learning, multiquery similarity-based models, and complex networks. The screening prioritized the antimicrobial activity while ensuring the absence of hemolytic and toxic properties, and structural uniqueness compared to known AMPs. Five relevant AMP datasets were released, ranging from a comprehensive collection of 542,485 AMPs to a refined dataset of 68,694 nonhemolytic and nontoxic AMPs. Further comparative analyses and application of network science principles helped identify 5466 unique and 808 representative nonhemolytic and nontoxic AMPs. These datasets, along with the selected mining tools, provide valuable resources for peptide drug developers.

In-silico and in-vitro study of novel antimicrobial peptide AM1 from Aegle marmelos against drug-resistant Staphylococcus aureus

Antimicrobial peptides have garnered increasing attention as potential alternatives due to their broad-spectrum antimicrobial activity and low propensity for developing resistance. This is for the first time; proteome sequences of Aegle marmelos were subjected to in-silico digestion and AMP prediction were performed using DBAASP server. After screening the peptides on the basis of different physiochemical property, peptide sequence GKEAATKAIKEWGQPKSKITH (AM1) shows the maximum binding affinity with - 10.2 Kcal/mol in comparison with the standard drug (Trimethoprim) with - 7.4 kcal/mol and - 6.8 Kcal/mol for DHFR and SaTrmK enzyme respectively. Molecular dynamics simulation performed for 300ns, it has been found that peptide was able to stabilize the protein more effectively, analysed by RMSD, RMSF, and other statistical analysis. Free binding energy for DHFR and SaTrmK interaction from MMPBSA analysis with peptide was found to be -47.69 and - 44.32 Kcal/mol and for Trimethoprim to be -13.85 Kcal/mol and - 11.67 Kcal/mol respectively. Further in-vitro study was performed against Methicillin Susceptible Staphylococcus aureus (MSSA), Methicillin Resistant Staphylococcus aureus (MRSA), Multi-Drug Resistant Staphylococcus aureus (MDR-SA) strain, where MIC values found to be 2, 4, and 8.5 µg/ml lesser in comparison to trimethoprim which has higher MIC values 2.5, 5, and 9.5 µg/ml respectively. Thus, our study provides the insight for the further in-vivo study of the peptides against multi-drug resistant S. aureus.

Peptide-Membrane Docking and Molecular Dynamic Simulation of In Silico Detected Antimicrobial Peptides from Portulaca oleracea's Transcriptome

The main issue with clinical infections is multidrug resistance to traditional antibiotics. As they are essential to innate immunity, shielding hosts from pathogenic microbes, traditional herbal remedies are an excellent supplier of antimicrobial peptides (AMPs), vital parts of defensive systems. Nevertheless, little is known about the bioactive peptide components of most ethnobotanical species. Our goal in this study was to find new, likely AMPs from Portulaca oleracea (P. oleracea) using in silico studies. The P. oleracea transcriptome was gained from Sequence Read Archive (SRA) and quality controlled, then adapters and other low-quality reads were trimmed. Afterward, de novo assembled and translated open reading frames (ORFs) were determined. Next, the ORFs were filtered based on AMP physiochemical criteria and deep learning methods. Finally, the five selected putative AMPs docked with E. coli and S. aureus membranes that showed penetration in bilayers. In this step, PO2 was chosen as a candidate AMP to analyze with molecular dynamics (MD) simulations. Our data demonstrated that PO2 is more stable in E. coli than in S. aureus. Moreover, these predicted AMPs can be good candidates for in vitro and in vivo analysis.

In silico directed evolution of Anabas testudineus AtMP1 antimicrobial peptide to improve in vitro anticancer activity

Various studies have demonstrated that directed evolution is a powerful tool in enhancing protein properties. In this study, directed evolution was used to enhance the efficacy of synthesised Anabas testudineus AtMP1 antimicrobial peptides (AMPs) in inhibiting the proliferation of cancer cells. The modification of antimicrobial peptides (AMPs) and prediction of peptide properties using bioinformatic tools were carried out using four databases, including ADP3, CAMP-R3, AMPfun, and ANTICP. One modified antimicrobial peptide (AMP), ATMP6 (THPPTTTTTTTTTTTTTAAPARTT), was chosen based on its projected potent anticancer effect, taking into account factors such as amino acid length, net charge, anticancer activity score, and hydrophobicity. The selected AMPs were subjected to study in deep-learning databases, namely ToxIBTL and ToxinPred2, to predict their toxicity. Furthermore, the allergic properties of these antimicrobial peptides (AMPs) were verified by utilising AllerTOP and AllergenFP. Based on the results obtained from the database study, it was projected that antimicrobial peptides (AMPs) demonstrate a lack of toxicity towards human cells that is indicative of the broader population. After 48 hours of incubation, the IC50 values of ATMP6 against the HS27 and MDA-MB-231 cell lines were found to be 48.03 ± 0.013 µg/ml and 7.52 ± 0.027 µg/ml, respectively. The IC50 values of the original peptide ATMP1 against the MDA-MB-231 and HS27 cell lines were determined to be 59.6 ± 0.14 µg/ml and 8.25 ± 0.14 µg/ml, respectively, when compared. Furthermore, the results indicated that the injection of ATMP6 induced apoptosis in the MDA-MB-231 cell lines. The present investigation has revealed new opportunities for advancing novel targeted peptide therapeutics to tackle cancer.

ToxGIN: an In silico prediction model for peptide toxicity via graph isomorphism networks integrating peptide sequence and structure information

Peptide drugs have demonstrated enormous potential in treating a variety of diseases, yet toxicity prediction remains a significant challenge in drug development. Existing models for prediction of peptide toxicity largely rely on sequence information and often neglect the three-dimensional (3D) structures of peptides. This study introduced a novel model for short peptide toxicity prediction, named ToxGIN. The model utilizes Graph Isomorphism Network (GIN), integrating the underlying amino acid sequence composition and the 3D structures of peptides. ToxGIN comprises three primary modules: (i) Sequence processing module, converting peptide 3D structures and sequences into information of nodes and edges; (ii) Feature extraction module, utilizing GIN to learn discriminative features from nodes and edges; (iii) Classification module, employing a fully connected classifier for toxicity prediction. ToxGIN performed well on the independent test set with F1 score = 0.83, AUROC = 0.91, and Matthews correlation coefficient = 0.68, better than existing models for prediction of peptide toxicity. These results validated the effectiveness of integrating 3D structural information with sequence data using GIN for peptide toxicity prediction. The proposed ToxGIN and data can be freely accessible at https://github.com/cihebiyql/ToxGIN.

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/).

Discovery of potential antidiabetic peptides using deep learning

Antidiabetic peptides (ADPs), peptides with potential antidiabetic activity, hold significant importance in the treatment and control of diabetes. Despite their therapeutic potential, the discovery and prediction of ADPs remain challenging due to limited data, the complex nature of peptide functions, and the expensive and time-consuming nature of traditional wet lab experiments. This study aims to address these challenges by exploring methods for the discovery and prediction of ADPs using advanced deep learning techniques. Specifically, we developed two models: a single-channel CNN and a three-channel neural network (CNN + RNN + Bi-LSTM). ADPs were primarily gathered from the BioDADPep database, alongside thousands of non-ADPs sourced from anticancer, antibacterial, and antiviral peptide datasets. Subsequently, data preprocessing was performed with the evolutionary scale model (ESM-2), followed by model training and evaluation through 10-fold cross-validation. Furthermore, this work collected a series of newly published ADPs as an independent test set through literature review, and found that the CNN model achieved the highest accuracy (90.48 %) in predicting the independent test set, surpassing existing ADP prediction tools. Finally, the application of the model was considered. SeqGAN was used to generate new candidate ADPs, followed by screening with the constructed CNN model. Selected peptides were then evaluated using physicochemical property prediction and structural forecasts for pharmaceutical potential. In summary, this study not only established robust ADP prediction models but also employed these models to screen a batch of potential ADPs, addressing a critical need in the field of peptide-based antidiabetic research.

Antimicrobial Peptide Identified via Machine Learning Presents Both Potent Antibacterial Properties and Low Toxicity toward Human Cells

Preventing infection is a critical clinical challenge; however, the extensive use of antibiotics has resulted in remarkably increased antibiotic resistance. A variety of antibiotic alternatives including antimicrobial peptides (AMPs) have been studied. Unfortunately, like most conventional antibiotics, most current AMPs have shown significantly high toxicity toward the host, and therefore induce compromised host responses that may lead to negative clinical outcomes such as delayed wound healing. In this study, one of the AMPs with a short length of nine amino acids was first identified via machine learning to present potentially low cytotoxicity, and then synthesized and validated in vitro against both bacteria and mammalian cells. It was found that this short AMP presented strong and fast-acting antimicrobial properties against bacteria like Staphylococcus aureus, one of the most common bacteria clinically, and it targeted and depolarized bacterial membranes. This AMP also demonstrated significantly lower (e.g., 30%) toxicity toward mammalian cells like osteoblasts, which are important cells for new bone formation, compared to conventional antibiotics like gentamicin, vancomycin, rifampin, cefazolin, and fusidic acid at short treatment times (e.g., 2 h). In addition, this short AMP demonstrated relatively low toxicity, similar to osteoblasts, toward an epithelial cell line like BEAS-2B cells.

Evaluation of quality consistency between dispensing granules and traditional decoction of Bombyx batryticatus based on peptidomics and in silico simulations

The application of traditional Chinese medicine dispensing granules is becoming increasingly prevalent. However, the consistency of dispensing granules with traditional decoction remains controversial. In this study, the consistency of peptide composition and pharmacodynamics between dispensing granules and traditional decoction of Bombyx batryticatus (BB) were assessed. A peptidomics method based on LC-tandem mass spectrometry technology was used to evaluate peptide composition similarity between BB traditional decoction and dispensing granules. The results revealed notable differences in peptide sequences between the two dosage forms, with only 8.55% of peptides shared between them. To evaluate the potential pharmacodynamic effects of the two dosage forms on epilepsy, virtual screening was used to identify potential active peptides, including blood-brain barrier permeability, toxicity prediction, and molecular docking. BB traditional decoction demonstrated a higher number and greater abundance of potential active peptides than BB dispensing granules, suggesting that BB traditional decoction may have a more favorable effect in treating epilepsy compared with BB dispensing granules. Moreover, molecular docking and molecular dynamics simulation studies confirmed the mechanism of action of active peptides to γ-aminobutyric acid transporter 1 (GAT-1). This study provides a scientific basis for the evaluation of quality consistency between BB traditional decoction and dispensing granules.

Identification and Design of Novel Potential Antimicrobial Peptides Targeting Mycobacterial Protein Kinase PknB

Antimicrobial peptides have gradually gained advantages over small molecule inhibitors for their multifunctional effects, synthesising accessibility and target specificity. The current study aims to determine an antimicrobial peptide to inhibit PknB, a serine/threonine protein kinase (STPK), by binding efficiently at the helically oriented hinge region. A library of 5626 antimicrobial peptides from publicly available repositories has been prepared and categorised based on the length. Molecular docking using ADCP helped to find the multiple conformations of the subjected peptides. For each peptide served as input the tool outputs 100 poses of the subjected peptide. To maintain an efficient binding for relatively a longer duration, only those peptides were chosen which were seen to bind constantly to the active site of the receptor protein over all the poses observed. Each peptide had different number of constituent amino acid residues; the peptides were classified based on the length into five groups. In each group the peptide length incremented upto four residues from the initial length form. Five peptides were selected for Molecular Dynamic simulation in Gromacs based on higher binding affinity. Post-dynamic analysis and the frame comparison inferred that neither the shorter nor the longer peptide but an intermediate length of 15 mer peptide bound well to the receptor. Residual substitution to the selected peptides was performed to enhance the targeted interaction. The new complexes considered were further analysed using the Elastic Network Model (ENM) for the functional site's intrinsic dynamic movement to estimate the new peptide's role. The study sheds light on prospects that besides the length of peptides, the combination of constituent residues equally plays a pivotal role in peptide-based inhibitor generation. The study envisages the challenges of fine-tuned peptide recovery and the scope of Machine Learning (ML) and Deep Learning (DL) algorithm development. As the study was primarily meant for generation of therapeutics for Tuberculosis (TB), the peptide proposed by this study demands meticulous invitro analysis prior to clinical applications.

Reverse vaccinology-based multi-epitope vaccine design against Indian group A rotavirus targeting VP7, VP4, and VP6 proteins

Rotavirus, a primary contributor to severe cases of infantile gastroenteritis on a global scale, results in significant morbidity and mortality in the under-five population, particularly in middle to low-income countries, including India. WHO-approved live-attenuated vaccines are linked to a heightened susceptibility to intussusception and exhibit low efficacy, primarily attributed to the high genetic diversity of rotavirus, varying over time and across different geographic regions. Herein, molecular data on Indian rotavirus A (RVA) has been reviewed through phylogenetic analysis, revealing G1P[8] to be the prevalent strain of RVA in India. The conserved capsid protein sequences of VP7, VP4 and VP6 were used to examine helper T lymphocyte, cytotoxic T lymphocyte and linear B-cell epitopes. Twenty epitopes were identified after evaluation of factors such as antigenicity, non-allergenicity, non-toxicity, and stability. These epitopes were then interconnected using suitable linkers and an N-terminal beta defensin adjuvant. The in silico designed vaccine exhibited structural stability and interactions with integrins (αvβ3 and αIIbβ3) and toll-like receptors (TLR2 and TLR4) indicated by docking and normal mode analyses. The immune simulation profile of the designed RVA multiepitope vaccine exhibited its potential to trigger humoral as well as cell-mediated immunity, indicating that it is a promising immunogen. These computational findings indicate potential efficacy of the designed vaccine against rotavirus infection.

ToxTeller: Predicting Peptide Toxicity Using Four Different Machine Learning Approaches

Examining the toxicity of peptides is essential for therapeutic peptide-based drug design. Machine learning approaches are frequently used to develop highly accurate predictors for peptide toxicity prediction. In this paper, we present ToxTeller, which provides four predictors using logistic regression, support vector machines, random forests, and XGBoost, respectively. For prediction model development, we construct a data set of toxic and nontoxic peptides from SwissProt and ConoServer databases with existence evidence levels checked. We also fully utilize the protein annotation in SwissProt to collect more toxic peptides than using keyword search alone. From this data set, we construct an independent test data set that shares at most 40% sequence similarity within itself and with the training data set. From a quite comprehensive list of 28 feature combinations, we conduct 10-fold cross-validation on the training data set to determine the optimized feature combination for model development. ToxTeller's performance is evaluated and compared with existing predictors on the independent test data set. Since toxic peptides must be avoided for drug design, we analyze strategies for reducing false-negative predictions of toxic peptides and suggest selecting models by top sensitivity instead of the widely used Matthews correlation coefficient, and also suggest using a meta-predictor approach with multiple predictors.

Structure-aware deep learning model for peptide toxicity prediction

Antimicrobial resistance is a critical public health concern, necessitating the exploration of alternative treatments. While antimicrobial peptides (AMPs) show promise, assessing their toxicity using traditional wet lab methods is both time-consuming and costly. We introduce tAMPer, a novel multi-modal deep learning model designed to predict peptide toxicity by integrating the underlying amino acid sequence composition and the three-dimensional structure of peptides. tAMPer adopts a graph-based representation for peptides, encoding ColabFold-predicted structures, where nodes represent amino acids and edges represent spatial interactions. Structural features are extracted using graph neural networks, and recurrent neural networks capture sequential dependencies. tAMPer's performance was assessed on a publicly available protein toxicity benchmark and an AMP hemolysis data we generated. On the latter, tAMPer achieves an F1-score of 68.7%, outperforming the second-best method by 23.4%. On the protein benchmark, tAMPer exhibited an improvement of over 3.0% in the F1-score compared to current state-of-the-art methods. We anticipate tAMPer to accelerate AMP discovery and development by reducing the reliance on laborious toxicity screening experiments.

T cell epitope based vaccine design while targeting outer capsid proteins of rotavirus strains infecting neonates: an immunoinformatics approach

Gastrointestinal diarrhea is majorly caused by the rotavirus (RV) in the children who generally are under the age group of 5 years. WHO estimates that ∼95% of the children contract RV infection, by this age. The disease is highly contagious; notably in many cases, it is proven fatal with high mortality rates especially in the developing countries. In India alone, an estimated 145,000 yearly deaths occurs due to RV related gastrointestinal diarrhea. WHO pre-qualified vaccines that are available for RV are all live attenuated vaccines with modest efficacy range between 40 and 60%. Further, the risk of intussusceptions has been reported in some children on RV vaccination. Thus, in a quest to develop alternative candidate to overcome challenges associated with these oral vaccines, we chose immunoinformatics approach to design a multi-epitope vaccine (MEV) while targeting the outer capsid viral proteinsVP4 and VP7 of the neonatal strains of rotavirus. Interestingly, ten epitopes, that is, six CD8+T-cells and four CD4+T-cell epitopes were identified which were predicted to be antigenic, non-allergic, non-toxic and stable. These epitopes were then linked to adjuvants, linkers, and PADRE sequences to create a multi-epitope vaccine for RV. The in silico designed RV-MEV and human TLR5 complex displayed stable interactions during molecular dynamics simulations. Further, the immune simulation studies of RV-MEV corroborated that the vaccine candidate emerges as a promising immunogen. Future investigations while performing in vitro and in vivo analyses with designed RV-MEV construct are highly desirable to warrant the potential of this vaccine candidate in protective immunity against different strains of RVs infecting neonates.Communicated by Ramaswamy H. Sarma.

The role and future prospects of artificial intelligence algorithms in peptide drug development

Peptide medications have been more well-known in recent years due to their many benefits, including low side effects, high biological activity, specificity, effectiveness, and so on. Over 100 peptide medications have been introduced to the market to treat a variety of illnesses. Most of these peptide medications are developed on the basis of endogenous peptides or natural peptides, which frequently required expensive, time-consuming, and extensive tests to confirm. As artificial intelligence advances quickly, it is now possible to build machine learning or deep learning models that screen a large number of candidate sequences for therapeutic peptides. Therapeutic peptides, such as those with antibacterial or anticancer properties, have been developed by the application of artificial intelligence algorithms.The process of finding and developing peptide drugs is outlined in this review, along with a few related cases that were helped by AI and conventional methods. These resources will open up new avenues for peptide drug development and discovery, helping to meet the pressing needs of clinical patients for disease treatment. Although peptide drugs are a new class of biopharmaceuticals that distinguish them from chemical and small molecule drugs, their clinical purpose and value cannot be ignored. However, the traditional peptide drug research and development has a long development cycle and high investment, and the creation of peptide medications will be substantially hastened by the AI-assisted (AI+) mode, offering a new boost for combating diseases.

VISH-Pred: an ensemble of fine-tuned ESM models for protein toxicity prediction

Peptide- and protein-based therapeutics are becoming a promising treatment regimen for myriad diseases. Toxicity of proteins is the primary hurdle for protein-based therapies. Thus, there is an urgent need for accurate in silico methods for determining toxic proteins to filter the pool of potential candidates. At the same time, it is imperative to precisely identify non-toxic proteins to expand the possibilities for protein-based biologics. To address this challenge, we proposed an ensemble framework, called VISH-Pred, comprising models built by fine-tuning ESM2 transformer models on a large, experimentally validated, curated dataset of protein and peptide toxicities. The primary steps in the VISH-Pred framework are to efficiently estimate protein toxicities taking just the protein sequence as input, employing an under sampling technique to handle the humongous class-imbalance in the data and learning representations from fine-tuned ESM2 protein language models which are then fed to machine learning techniques such as Lightgbm and XGBoost. The VISH-Pred framework is able to correctly identify both peptides/proteins with potential toxicity and non-toxic proteins, achieving a Matthews correlation coefficient of 0.737, 0.716 and 0.322 and F1-score of 0.759, 0.696 and 0.713 on three non-redundant blind tests, respectively, outperforming other methods by over $10\%$ on these quality metrics. Moreover, VISH-Pred achieved the best accuracy and area under receiver operating curve scores on these independent test sets, highlighting the robustness and generalization capability of the framework. By making VISH-Pred available as an easy-to-use web server, we expect it to serve as a valuable asset for future endeavors aimed at discerning the toxicity of peptides and enabling efficient protein-based therapeutics.

Introducing enzymatic cleavage features and transfer learning realizes accurate peptide half-life prediction across species and organs

Peptide drugs are becoming star drug agents with high efficiency and selectivity which open up new therapeutic avenues for various diseases. However, the sensitivity to hydrolase and the relatively short half-life have severely hindered their development. In this study, a new generation artificial intelligence-based system for accurate prediction of peptide half-life was proposed, which realized the half-life prediction of both natural and modified peptides and successfully bridged the evaluation possibility between two important species (human, mouse) and two organs (blood, intestine). To achieve this, enzymatic cleavage descriptors were integrated with traditional peptide descriptors to construct a better representation. Then, robust models with accurate performance were established by comparing traditional machine learning and transfer learning, systematically. Results indicated that enzymatic cleavage features could certainly enhance model performance. The deep learning model integrating transfer learning significantly improved predictive accuracy, achieving remarkable R2 values: 0.84 for natural peptides and 0.90 for modified peptides in human blood, 0.984 for natural peptides and 0.93 for modified peptides in mouse blood, and 0.94 for modified peptides in mouse intestine on the test set, respectively. These models not only successfully composed the above-mentioned system but also improved by approximately 15% in terms of correlation compared to related works. This study is expected to provide powerful solutions for peptide half-life evaluation and boost peptide drug development.

Mechanism Analysis of Antimicrobial Peptide NoPv1 Related to Potato Late Blight through a Computer-Aided Study

Phytophthora infestans (Mont.) de Bary, the oomycotic pathogen responsible for potato late blight, is the most devastating disease of potato production. The primary pesticides used to control oomycosis are phenyl amide fungicides, which cause environmental pollution and toxic residues harmful to both human and animal health. To address this, an antimicrobial peptide, NoPv1, has been screened to target Plasmopara viticola cellulose synthase 2 (PvCesA2) to inhibit the growth of Phytophthora infestans (P. infestans). In this study, we employed AlphaFold2 to predict the three-dimensional structure of PvCesA2 along with NoPv peptides. Subsequently, utilizing computational methods, we dissected the interaction mechanism between PvCesA2 and these peptides. Based on this analysis, we performed a saturation mutation of NoPv1 and successfully obtained the double mutants DP1 and DP2 with a higher affinity for PvCesA2. Meanwhile, dynamics simulations revealed that both DP1 and DP2 utilize a mechanism akin to the barrel-stave model for penetrating the cell membrane. Furthermore, the predicted results showed that the antimicrobial activity of DP1 was superior to that of NoPv1 without being toxic to human cells. These findings may offer insights for advancing the development of eco-friendly pesticides targeting various oomycete diseases, including late blight.

Integrated convolution and self-attention for improving peptide toxicity prediction

MOTIVATION

Peptides are promising agents for the treatment of a variety of diseases due to their specificity and efficacy. However, the development of peptide-based drugs is often hindered by the potential toxicity of peptides, which poses a significant barrier to their clinical application. Traditional experimental methods for evaluating peptide toxicity are time-consuming and costly, making the development process inefficient. Therefore, there is an urgent need for computational tools specifically designed to predict peptide toxicity accurately and rapidly, facilitating the identification of safe peptide candidates for drug development.

RESULTS

We provide here a novel computational approach, CAPTP, which leverages the power of convolutional and self-attention to enhance the prediction of peptide toxicity from amino acid sequences. CAPTP demonstrates outstanding performance, achieving a Matthews correlation coefficient of approximately 0.82 in both cross-validation settings and on independent test datasets. This performance surpasses that of existing state-of-the-art peptide toxicity predictors. Importantly, CAPTP maintains its robustness and generalizability even when dealing with data imbalances. Further analysis by CAPTP reveals that certain sequential patterns, particularly in the head and central regions of peptides, are crucial in determining their toxicity. This insight can significantly inform and guide the design of safer peptide drugs.

AVAILABILITY AND IMPLEMENTATION

The source code for CAPTP is freely available at https://github.com/jiaoshihu/CAPTP.

Machine learning for antimicrobial peptide identification and design

Artificial intelligence (AI) and machine learning (ML) models are being deployed in many domains of society and have recently reached the field of drug discovery. Given the increasing prevalence of antimicrobial resistance, as well as the challenges intrinsic to antibiotic development, there is an urgent need to accelerate the design of new antimicrobial therapies. Antimicrobial peptides (AMPs) are therapeutic agents for treating bacterial infections, but their translation into the clinic has been slow owing to toxicity, poor stability, limited cellular penetration and high cost, among other issues. Recent advances in AI and ML have led to breakthroughs in our abilities to predict biomolecular properties and structures and to generate new molecules. The ML-based modelling of peptides may overcome some of the disadvantages associated with traditional drug discovery and aid the rapid development and translation of AMPs. Here, we provide an introduction to this emerging field and survey ML approaches that can be used to address issues currently hindering AMP development. We also outline important limitations that can be addressed for the broader adoption of AMPs in clinical practice, as well as new opportunities in data-driven peptide design.

Toxicity prediction and classification of Gunqile-7 with small sample based on transfer learning method

Drug-induced diseases are the most important component of iatrogenic disease. It is the duty of doctors to provide a reasonable and safe dose of medication. Gunqile-7 is a Mongolian medicine with analgesic and anti-inflammatory effects. As a foreign substance in the body, even with reasonable medication, it may produce varying degrees of adverse reactions or toxic side effects. Since the cost of collecting Gunqile-7 for pharmacological animal trials is high and the data sample is small, this paper employs transfer learning and data augmentation methods to study the toxicity of Gunqile-7. More specifically, to reduce the necessary number of training samples, the data augmentation approach is employed to extend the data set. Then, the transfer learning method and one-dimensional convolutional neural network are utilized to train the network. In addition, we use the support vector machine-recursive feature elimination method for feature selection to reduce features that have adverse effects on model predictions. Furthermore, due to the important role of the pre-trained model of transfer learning, we select a quantitative toxicity prediction model as the pre-trained model, which is consistent with the purpose of this paper. Lastly, the experimental results demonstrate the efficiency of the proposed method. Our method can improve accuracy by up to 9 percentage points compared to the method without transfer learning on a small sample set.

MultiToxPred 1.0: a novel comprehensive tool for predicting 27 classes of protein toxins using an ensemble machine learning approach

Protein toxins are defense mechanisms and adaptations found in various organisms and microorganisms, and their use in scientific research as therapeutic candidates is gaining relevance due to their effectiveness and specificity against cellular targets. However, discovering these toxins is time-consuming and expensive. In silico tools, particularly those based on machine learning and deep learning, have emerged as valuable resources to address this challenge. Existing tools primarily focus on binary classification, determining whether a protein is a toxin or not, and occasionally identifying specific types of toxins. For the first time, we propose a novel approach capable of classifying protein toxins into 27 distinct categories based on their mode of action within cells. To accomplish this, we assessed multiple machine learning techniques and found that an ensemble model incorporating the Light Gradient Boosting Machine and Quadratic Discriminant Analysis algorithms exhibited the best performance. During the tenfold cross-validation on the training dataset, our model exhibited notable metrics: 0.840 accuracy, 0.827 F1 score, 0.836 precision, 0.840 sensitivity, and 0.989 AUC. In the testing stage, using an independent dataset, the model achieved 0.846 accuracy, 0.838 F1 score, 0.847 precision, 0.849 sensitivity, and 0.991 AUC. These results present a powerful next-generation tool called MultiToxPred 1.0, accessible through a web application. We believe that MultiToxPred 1.0 has the potential to become an indispensable resource for researchers, facilitating the efficient identification of protein toxins. By leveraging this tool, scientists can accelerate their search for these toxins and advance their understanding of their therapeutic potential.

Novel Antimicrobial Peptide SAAP Mutant as a Better Adjuvant to Sulbactam-Based Treatments Against Clinical Strains of XDR Acinetobacter baumannii

The production of extended spectrum β-lactamases (ESBLs) in extensively drug-resistant (XDR) strains of Acinetobacter baumannii has created havoc amongst clinicians making the treatment procedure challenging. Carbapenem-resistant strains have displayed total ineffectiveness towards newer combinations of β-lactam-β-lactamase inhibitors (βL-βLI) in tertiary healthcare settings. Therefore, the present study was aimed to design potential β-lactamase antimicrobial peptide (AMP) inhibitors against ESBLs produced by the strains. We have constructed an AMP mutant library with higher antimicrobial efficacy (range: ~ 15 to 27%) than their parent peptides. The mutants were thoroughly screened based on different physicochemical and immunogenic properties revealing three peptides, namely SAAP-148, HFIAP-1, myticalin-C6 and their mutants with safe pharmacokinetics profile. Molecular docking highlighted SAAP-148_M15 displaying maximum inhibitory potential with lowest binding energies against NDM1 (- 1148.7 kcal/mol), followed by OXA23 (- 1032.5 kcal/mol) and OXA58 (- 925.3 kcal/mol). The intermolecular interaction profiles displayed SAAP-148_M15 exhibiting hydrogen bonds and van der Waals hydrophobic interactions with the crucial residues of metallo β-lactamase [IPR001279] and penicillin-binding transpeptidase [IPR001460] domains. Coarse-grained clustering and molecular dynamics simulations (MDS) further validated the stable backbone profile and minimal residue-level fluctuations of the protein-peptide complex that were maintained throughout the simulation timeframe. The present study hypothesised that the combination of sulbactam (βL) with SAAP-148_M15 (βLI) holds immense potential in inhibiting the ESBLs alongside restoration of sulbactam activity. The current in silico findings upon further experimental validations can pave path towards designing of successful therapeutic strategy against XDR strains of A. baumannii.

Immunoinformatics-Driven Strategies for Advancing Epitope-Based Vaccine Design for West Nile Virus

The West Nile virus (WNV) is the causative agent of West Nile disease (WND), which poses a potential risk of meningitis or encephalitis. The aim of the study was to design an epitope-based vaccine for WNV by utilizing computational analyses. The epitope-based vaccine design process encompassed WNV sequence collection, phylogenetic tree construction, and sequence alignment. Computational models identified B-cell and T-cell epitopes, followed by immunological property analysis. Epitopes were then modeled and docked with B-cell receptors, MHC I, and MHC II. Molecular dynamics simulations further explored dynamic interactions between epitopes and receptors. The findings indicated that the B-cell epitope QINHHWHKSGSSIG, along with three T-cell epitopes (FLVHREWFM for MHC I, NPFVSVATANAKVLI for MHC II, and NAYYVMTVGTKTFLV for MHC II), successfully passed the immunological evaluations. These four epitopes were further subjected to docking and molecular dynamics simulation studies. Although each demonstrated favorable affinities with their respective receptors, only NAYYVMTVGTKTFLV displayed a stable interaction with MHC II during MDS analysis, hence emerging as a potential candidate for a WNV epitope-based vaccine. This study demonstrates a comprehensive approach to epitope vaccine design, combining computational analyses, molecular modeling, and simulation techniques to identify potential vaccine candidates for WNV.

De novo antioxidant peptide design via machine learning and DFT studies

Antioxidant peptides (AOPs) are highly valued in food and pharmaceutical industries due to their significant role in human function. This study introduces a novel approach to identifying robust AOPs using a deep generative model based on sequence representation. Through filtration with a deep-learning classification model and subsequent clustering via the Butina cluster algorithm, twelve peptides (GP1-GP12) with potential antioxidant capacity were predicted. Density functional theory (DFT) calculations guided the selection of six peptides for synthesis and biological experiments. Molecular orbital representations revealed that the HOMO for these peptides is primarily localized on the indole segment, underscoring its pivotal role in antioxidant activity. All six synthesized peptides exhibited antioxidant activity in the DPPH assay, while the hydroxyl radical test showed suboptimal results. A hemolysis assay confirmed the non-hemolytic nature of the generated peptides. Additionally, an in silico investigation explored the potential inhibitory interaction between the peptides and the Keap1 protein. Analysis revealed that ligands GP3, GP4, and GP12 induced significant structural changes in proteins, affecting their stability and flexibility. These findings highlight the capability of machine learning approaches in generating novel antioxidant peptides.

Dual-View Learning Based on Images and Sequences for Molecular Property Prediction

The prediction of molecular properties remains a challenging task in the field of drug design and development. Recently, there has been a growing interest in the analysis of biological images. Molecular images, as a novel representation, have proven to be competitive, yet they lack explicit information and detailed semantic richness. Conversely, semantic information in SMILES sequences is explicit but lacks spatial structural details. Therefore, in this study, we focus on and explore the relationship between these two types of representations, proposing a novel multimodal architecture named ISMol. ISMol relies on a cross-attention mechanism to extract information representations of molecules from both images and SMILES strings, thereby predicting molecular properties. Evaluation results on 14 small molecule ADMET datasets indicate that ISMol outperforms machine learning (ML) and deep learning (DL) models based on single-modal representations. In addition, we analyze our method through a large number of experiments to test the superiority, interpretability and generalizability of the method. In summary, ISMol offers a powerful deep learning toolbox for drug discovery in a variety of molecular properties.

Inactivation of cell-free HIV-1 by designing potent peptides based on mutations in the CD4 binding site

Human immunodeficiency virus type 1 (HIV-1) is a major global health problem, with over 38 million people infected worldwide. Current anti-HIV-1 drugs are limited in their ability to prevent the virus from replicating inside host cells, making them less effective as preventive measures. In contrast, viral inhibitors that inactivate the virus before it can bind to a host cell have great potential as drugs. In this study, we aimed to design mutant peptides that could block the interaction between gp120 and the CD4 receptor on host cells, thus preventing HIV-1 infection. We designed a 20-amino-acid peptide that mimicked the amino acids of the CD4 binding site and docked it to gp120. Molecular dynamics simulations were performed to calculate the energy of MMPBSA (Poisson-Boltzmann Surface Area) for each residue of the peptide, and unfavorable energy residues were identified as potential mutation points. Using MAESTRO (Multi AgEnt STability pRedictiOn), we measured ΔΔG (change in the change in Gibbs free energy) for mutations and generated a library of 240 mutated peptides using OSPREY software. The peptides were then screened for allergenicity and binding affinity. Finally, molecular dynamics simulations (via GROMACS 2020.2) and control docking (via HADDOCK 2.4) were used to evaluate the ability of four selected peptides to inhibit HIV-1 infection. Three peptides, P3 (AHRQIRQWFLTRGPNRSLWQ), P4 (VHRQIRQWFLTRGPNRSLWQ), and P9 (AHRQIRQMFLTRGPNRSLWQ), showed practical and potential as HIV inhibitors, based on their binding affinity and ability to inhibit infection. These peptides have the ability to inactivate the virus before it can bind to a host cell, thus representing a promising approach to HIV-1 prevention. Our findings suggest that mutant peptides designed to block the interaction between gp120 and the CD4 receptor have potential as HIV-1 inhibitors. These peptides could be used as preventive measures against HIV-1 transmission, and further research is needed to evaluate their safety and efficacy in clinical settings.

Invited review: Camel milk-derived bioactive peptides and diabetes-Molecular view and perspectives

In dairy science, camel milk (CM) constitutes a center of interest for scientists due to its known beneficial effect on diabetes as demonstrated in many in vitro, in vivo, and clinical studies and trials. Overall, CM had positive effects on various parameters related to glucose transport and metabolism as well as the structural and functional properties of the pancreatic β-cells and insulin secretion. Thus, CM consumption may help manage diabetes; however, such a recommendation will become rationale and clinically conceivable only if the exact molecular mechanisms and pathways involved at the cellular levels are well understood. Moreover, the application of CM as an alternative antidiabetic tool may first require the identification of the exact bioactive molecules behind such antidiabetic properties. In this review, we describe the advances in our knowledge of the molecular mechanisms reported to be involved in the beneficial effects of CM in managing diabetes using different in vitro and in vivo models. This mainly includes the effects of CM on the different molecular pathways controlling (1) insulin receptor signaling and glucose uptake, (2) the pancreatic β-cell structure and function, and (3) the activity of key metabolic enzymes in glucose metabolism. Moreover, we described the current status of the identification of CM-derived bioactive peptides and their structure-activity relationship study and characterization in the context of molecular markers related to diabetes. Such an overview will not only enrich our scientific knowledge of the plausible mode of action of CM in diabetes but should ultimately rationalize the claim of the potential application of CM against diabetes. This will pave the way toward new directions and ideas for developing a new generation of antidiabetic products taking benefits from the chemical composition of CM.

Multi-CGAN: Deep Generative Model-Based Multiproperty Antimicrobial Peptide Design

Antimicrobial peptides are peptides that are effective against bacteria and viruses, and the discovery of new antimicrobial peptides is of great importance to human life and health. Although the design of antimicrobial peptides using machine learning methods has achieved good results in recent years, it remains a challenge to learn and design novel antimicrobial peptides with multiple properties of interest from peptide data with certain property labels. To this end, we propose Multi-CGAN, a deep generative model-based architecture that can learn from single-attribute peptide data and generate antimicrobial peptide sequences with multiple attributes that we need, which may have a potentially wide range of uses in drug discovery. In particular, we verified that our Multi-CGAN generated peptides with the desired properties have good performance in terms of generation rate. Moreover, a comprehensive statistical analysis demonstrated that our generated peptides are diverse and have a low probability of being homologous to the training data. Interestingly, we found that the performance of many popular deep learning methods on the antimicrobial peptide prediction task can be improved by using Multi-CGAN to expand the data on the training set of the original task, indicating the high quality of our generated peptides and the robust ability of our method. In addition, we also investigated whether it is possible to directionally generate peptide sequences with specified properties by controlling the input noise sampling for our model.

Computational identification of a multi-peptide vaccine candidate in E2 glycoprotein against diverse Hepatitis C virus genotypes

Hepatitis C Virus (HCV) is estimated to affect nearly 180 million people worldwide, culminating in ∼0.7 million yearly casualties. However, a safe vaccine against HCV is not yet available. This study endeavored to identify a multi-genotypic, multi-epitopic, safe, and globally competent HCV vaccine candidate. We employed a consensus epitope prediction strategy to identify multi-epitopic peptides in all known envelope glycoprotein (E2) sequences, belonging to diverse HCV genotypes. The obtained peptides were screened for toxicity, allergenicity, autoimmunity and antigenicity, resulting in two favorable peptides viz., P2 (VYCFTPSPVVVG) and P3 (YRLWHYPCTV). Evolutionary conservation analysis indicated that P2 and P3 are highly conserved, supporting their use as part of a designed multi-genotypic vaccine. Population coverage analysis revealed that P2 and P3 are likely to be presented by >89% Human Leukocyte Antigen (HLA) molecules from six geographical regions. Indeed, molecular docking predicted the physical binding of P2 and P3 to various representative HLAs. We designed a vaccine construct using these peptides and assessed its binding to toll-like receptor 4 (TLR-4) by molecular docking and simulation. Subsequent analysis by energy-based and machine learning tools predicted high binding affinity and pinpointed the key binding residues (i.e. hotspots) in P2 and P3. Also, a favorable immunogenic profile of the construct was predicted by immune simulations. We encourage the scientific community to validate our vaccine construct in vitro and in vivo.Communicated by Ramaswamy H. Sarma.

IIFS: An improved incremental feature selection method for protein sequence processing

MOTIVATION

Discrete features can be obtained from protein sequences using a feature extraction method. These features are the basis of downstream processing of protein data, but it is necessary to screen and select some important features from them as they generally have data redundancy.

RESULT

Here, we report IIFS, an improved incremental feature selection method that exploits a new subset search strategy to find the optimal feature set. IIFS combines nonadjacent sorting features to prevent the drawbacks of data explosion and excessive reliance on feature sorting results. The comparative experimental results on 27 feature sorting data show that IIFS can find more accurate and important features compared to existing methods.The IIFS approach also handles data redundancy more efficiently and finds more representative and discriminatory features while ensuring minimal feature dimensionality and good evaluation metrics. Moreover, we wrap this method and deploy it on a web server for access at http://112.124.26.17:8005/.

aSynPEP-DB: a database of biogenic peptides for inhibiting α-synuclein aggregation

Parkinson's disease (PD) is the second most prevalent neurodegenerative disorder, yet effective treatments able to stop or delay disease progression remain elusive. The aggregation of a presynaptic protein, α-synuclein (aSyn), is the primary neurological hallmark of PD and, thus, a promising target for therapeutic intervention. However, the lack of consensus on the molecular properties required to specifically bind the toxic species formed during aSyn aggregation has hindered the development of therapeutic molecules. Recently, we defined and experimentally validated a peptide architecture that demonstrated high affinity and selectivity in binding to aSyn toxic oligomers and fibrils, effectively preventing aSyn pathogenic aggregation. Human peptides with such properties may have neuroprotective activities and hold a huge therapeutic interest. Driven by this idea, here, we developed a discriminative algorithm for the screening of human endogenous neuropeptides, antimicrobial peptides and diet-derived bioactive peptides with the potential to inhibit aSyn aggregation. We identified over 100 unique biogenic peptide candidates and ensembled a comprehensive database (aSynPEP-DB) that collects their physicochemical features, source datasets and additional therapeutic-relevant information, including their sites of expression and associated pathways. Besides, we provide access to the discriminative algorithm to extend its application to the screening of artificial peptides or new peptide datasets. aSynPEP-DB is a unique repository of peptides with the potential to modulate aSyn aggregation, serving as a platform for the identification of previously unexplored therapeutic agents. Database URL:  https://asynpepdb.ppmclab.com/.

DeepAFP: An effective computational framework for identifying antifungal peptides based on deep learning

Fungal infections have become a significant global health issue, affecting millions worldwide. Antifungal peptides (AFPs) have emerged as a promising alternative to conventional antifungal drugs due to their low toxicity and low propensity for inducing resistance. In this study, we developed a deep learning-based framework called DeepAFP to efficiently identify AFPs. DeepAFP fully leverages and mines composition information, evolutionary information, and physicochemical properties of peptides by employing combined kernels from multiple branches of convolutional neural network with bi-directional long short-term memory layers. In addition, DeepAFP integrates a transfer learning strategy to obtain efficient representations of peptides for improving model performance. DeepAFP demonstrates strong predictive ability on carefully curated datasets, yielding an accuracy of 93.29% and an F1-score of 93.45% on the DeepAFP-Main dataset. The experimental results show that DeepAFP outperforms existing AFP prediction tools, achieving state-of-the-art performance. Finally, we provide a downloadable AFP prediction tool to meet the demands of large-scale prediction and facilitate the usage of our framework by the public or other researchers. Our framework can accurately identify AFPs in a short time without requiring significant human and material resources, and hence can accelerate the development of AFPs as well as contribute to the treatment of fungal infections. Furthermore, our method can provide new perspectives for other biological sequence analysis tasks.

Aquatic Peptide: The Potential Anti-Cancer and Anti-Microbial Activity of GE18 Derived from Pathogenic Fungus Aphanomyces invadans

Small molecules as well as peptide-based therapeutic approaches have attracted global interest due to their lower or no toxicity in nature, and their potential in addressing several health complications including immune diseases, cardiovascular diseases, metabolic disorders, osteoporosis and cancer. This study proposed a peptide, GE18 of subtilisin-like peptidase from the virulence factor of aquatic pathogenic fungus Aphanomyces invadans, which elicits anti-cancer and anti-microbial activities. To understand the potential GE18 peptide-induced biological effects, an in silico analysis, in vitro (L6 cells) and in vivo toxicity assays (using zebrafish embryo), in vitro anti-cancer assays and anti-microbial assays were performed. The outcomes of the in silico analyses demonstrated that the GE18 peptide has potent anti-cancer and anti-microbial activities. GE18 is non-toxic to in vitro non-cancerous cells and in vivo zebrafish larvae. However, the peptide showed significant anti-cancer properties against MCF-7 cells with an IC50 value of 35.34 µM, at 24 h. Besides the anti-proliferative effect on cancer cells, the peptide exposure does promote the ROS concentration, mitochondrial membrane potential and the subsequent upregulation of anti-cancer genes. On the other hand, GE18 elicits significant anti-microbial activity against P. aeruginosa, wherein GE18 significantly inhibits bacterial biofilm formation. Since the peptide has positively charged amino acid residues, it targets the cell membrane, as is evident in the FESEM analysis. Based on these outcomes, it is possible that the GE18 peptide is a significant anti-cancer and anti-microbial molecule.