Detoxification and underlying mechanisms towards toxic alkaloids by Traditional Chinese Medicine processing: A comprehensive review.

BACKGROUND: Alkaloids have attracted enduring interest worldwide due to their remarkable therapeutic effects, including analgesic, anti-inflammatory, and anti-tumor properties, thus offering a rich source for lead compound design and new drug discovery. However, some of these alkaloids possess intrinsic toxicity. Processing (Paozhi) is a pre-treatment step before the application of herbal medicines in traditional Chinese medicine (TCM) clinics, which has been employed for centuries to mitigate the toxicity of alkaloid-rich TCMs. PURPOSE: To explore the toxicity phenotypes, chemical basis, mode of action, detoxification processing methods, and underlying mechanisms, we can gain crucial insights into the safe and rational use of these toxic alkaloid-rich herbs. Such insights have the great potential to offer new strategies for drug discovery and development, ultimately improving the quality of life for millions of people. METHODS: Literatures published or early accessed until December 31, 2023, were retrieved from databases including PubMed, Web of Science, and CNKI. The following keywords, such as "toxicity", "alkaloid", "detoxification", "processing", "traditional Chinese medicine", "medicinal plant", and "plant", were used in combination or separately for screening. RESULTS: Toxicity of alkaloids in TCM includes hepatotoxicity, nephrotoxicity, neurotoxicity, cardiotoxicity, and other forms of toxicity, primarily induced by pyrrolizidines, quinolizidines, isoquinolines, indoles, pyridines, terpenoids, and amines. Factors such as whether the toxic-alkaloid enriched part is limited or heat-sensitive, and whether toxic alkaloids are also therapeutic components, are critical for choosing appropriate detoxification processing methods. Mechanisms of alkaloid detoxification includes physical removal, chemical decomposition or transformation, as well as biological modifications. CONCLUSION: Through this exploration, we review toxic alkaloids and the mechanisms underlying their toxicity, discuss methods to reduce toxicity, and unravel the intricate mechanisms behind detoxification. These offers insights into the quality control of herbs containing toxic alkaloids, safe and rational use of alkaloid-rich TCMs in clinics, new strategies for drug discovery and development, and ultimately helping improve the quality of life for millions of people.

MultiModRLBP: A Deep Learning Approach for Multi-Modal RNA-Small Molecule Ligand Binding Sites Prediction.

This study aims to tackle the intricate challenge of predicting RNA-small molecule binding sites to explore the potential value in the field of RNA drug targets. To address this challenge, we propose the MultiModRLBP method, which integrates multi-modal features using deep learning algorithms. These features include 3D structural properties at the nucleotide base level of the RNA molecule, relational graphs based on overall RNA structure, and rich RNA semantic information. In our investigation, we gathered 851 interactions between RNA and small molecule ligand from the RNAglib dataset and RLBind training set. Unlike conventional training sets, this collection broadened its scope by including RNA complexes that have the same RNA sequence but change their respective binding sites due to structural differences or the presence of different ligands. This enhancement enables the MultiModRLBP model to more accurately capture subtle changes at the structural level, ultimately improving its ability to discern nuances among similar RNA conformations. Furthermore, we evaluated MultiModRLBP on two classic test sets, Test18 and Test3, highlighting its performance disparities on small molecules based on metal and non-metal ions. Additionally, we conducted a structural sensitivity analysis on specific complex categories, considering RNA instances with varying degrees of structural changes and whether they share the same ligands. The research results indicate that MultiModRLBP outperforms the current state-of-the-art methods on multiple classic test sets, particularly excelling in predicting binding sites for non-metal ions and instances where the binding sites are widely distributed along the sequence. MultiModRLBP also can be used as a potential tool when the RNA structure is perturbed or the RNA experimental tertiary structure is not available. Most importantly, MultiModRLBP exhibits the capability to distinguish binding characteristics of RNA that are structurally diverse yet exhibit sequence similarity. These advancements hold promise in reducing the costs associated with the development of RNA-targeted drugs.

Enhanced water treatment performance of ceramic-based forward osmosis membranes via MOF interlayer.

Forward osmosis (FO) membrane processes could operate without hydraulic pressures, enabling the efficient treatment of wastewaters with mitigated membrane fouling and enhanced efficiency. Designing a high-performance polyamide (PA) layer on ceramic substrates remains a challenge for FO desalination applications. Herein, we report the enhanced water treatment performance of thin-film nanocomposite ceramic-based FO membranes via an in situ grown Zr-MOF (UiO-66-NH2) interlayer. With the Zr-MOF interlayer, the ceramic-based FO membranes exhibit lower thickness, higher cross-linking degree, and increased surface roughness, leading to higher water flux of 27.38 L m-2 h-1 and lower reverse salt flux of 3.45 g m-2 h-1. The ceramic-based FO membranes with Zr-MOF interlayer not only have an application potential in harsh environments such as acidic solution (pH 3) and alkaline solution (pH 11), but also exhibit promising water and reverse salt transport properties, which are better than most MOF-incorporated PA membranes. Furthermore, the membranes could reject major species (ions, oil and organics) with rejections ＞94 % and water flux of 22.62-14.35 L m-2 h-1 in the treatment of actual alkaline industrial wastewater (pH 8.6). This rational design proposed in this study is not only applicable for the development of a high-quality ceramic-based FO membrane with enhanced performance but also can be potentially extended to more challenging water treatment applications.

Bifidobacterium alleviate metabolic disorders via converting methionine to 5'-methylthioadenosine.

Dietary patterns and corresponding gut microbiota profiles are associated with various health conditions. A diet rich in polyphenols, primarily plant-based, has been shown to promote the growth of probiotic bacteria in the gastrointestinal tract, subsequently reducing the risk of metabolic disorders in the host. The beneficial effects of these bacteria are largely due to the specific metabolites they produce, such as short-chain fatty acids and membrane proteins. In this study, we employed a metabolomics-guided bioactive metabolite identification platform that included bioactivity testing using in vitro and in vivo assays to discover a bioactive metabolite produced from probiotic bacteria. Through this approach, we identified 5'-methylthioadenosine (MTA) as a probiotic bacterial-derived metabolite with anti-obesity properties. Furthermore, our findings indicate that MTA administration has several regulatory impacts on liver functions, including modulating fatty acid synthesis and glucose metabolism. The present study elucidates the intricate interplay between dietary habits, gut microbiota, and their resultant metabolites.

Proteomic analysis of multiple organ dysfunction induced by rhabdomyolysis.

Rhabdomyolysis (RM) leads to dysfunction in the core organs of kidney, lung and heart, which is an important reason for the high mortality and disability rate of this disease. However, there is a lack of systematic research on the characteristics of rhabdomyolysis-induced injury in various organs and the underlying pathogenetic mechanisms, and especially the interaction between organs. We established a rhabdomyolysis model, observed the structural and functional changes in kidney, heart, and lung. It is observed that rhabdomyolysis results in significant damage in kidney, lung and heart of rats, among which the pathological damage of kidney and lung was significant, and of heart was relatively light. Meanwhile, we analyzed the differentially expressed proteins (DEPs) in the kidney, heart and lung between the RM group and the sham group based on liquid chromatography-tandem mass spectrometry (LC-MS/MS). In our study, Serpina3n was significantly up-regulated in the kidney, heart and lung. Serpina3n is a secreted protein and specifically inhibits a variety of proteases and participates in multiple physiological processes such as complement activation, inflammatory responses, apoptosis pathways, and extracellular matrix metabolism. It is inferred that Serpina3n may play an important role in multiple organ damage caused by rhabdomyolysis and could be used as a potential biomarker. This study comprehensively describes the functional and structural changes of kidney, heart and lung in rats after rhabdomyolysis, analyzes the DEPs of kidney, heart and lung, and determines the key role of Serpina3n in multiple organ injury caused by rhabdomyolysis. SIGNIFICANCE: This study comprehensively describes the functional and structural changes of kidney, heart and lung in rats after rhabdomyolysis, analyzes the DEPs of kidney, heart and lung, and determines the key role of Serpina3n in multiple organ injury caused by rhabdomyolysis.

RPEMHC: improved prediction of MHC-peptide binding affinity by a deep learning approach based on residue-residue pair encoding.

MOTIVATION: Binding of peptides to major histocompatibility complex (MHC) molecules plays a crucial role in triggering T cell recognition mechanisms essential for immune response. Accurate prediction of MHC-peptide binding is vital for the development of cancer therapeutic vaccines. While recent deep learning-based methods have achieved significant performance in predicting MHC-peptide binding affinity, most of them separately encode MHC molecules and peptides as inputs, potentially overlooking critical interaction information between the two. RESULTS: In this work, we propose RPEMHC, a new deep learning approach based on residue-residue pair encoding to predict the binding affinity between peptides and MHC, which encode an MHC molecule and a peptide as a residue-residue pair map. We evaluate the performance of RPEMHC on various MHC-II-related datasets for MHC-peptide binding prediction, demonstrating that RPEMHC achieves better or comparable performance against other state-of-the-art baselines. Moreover, we further construct experiments on MHC-I-related datasets, and experimental results demonstrate that our method can work on both two MHC classes. These extensive validations have manifested that RPEMHC is an effective tool for studying MHC-peptide interactions and can potentially facilitate the vaccine development. AVAILABILITY: The source code of the method along with trained models is freely available at https://github.com/lennylv/RPEMHC.

Intermittent short-duration reoxygenation relieves high-altitude pulmonary hypertension via NOX4/H2O2/PPAR-γ axis.

High-altitude pulmonary hypertension (HAPH) is a severe and progressive disease that can lead to right heart failure. Intermittent short-duration reoxygenation at high altitude is effective in alleviating HAPH; however, the underlying mechanisms are unclear. In the present study, a simulated 5,000-m hypoxia rat model and hypoxic cultured pulmonary artery smooth muscle cells (PASMCs) were used to evaluate the effect and mechanisms of intermittent short-duration reoxygenation. The results showed that intermittent 3-h/per day reoxygenation (I3) effectively attenuated chronic hypoxia-induced pulmonary hypertension and reduced the content of H2O2 and the expression of NADPH oxidase 4 (NOX4) in lung tissues. In combination with I3, while the NOX inhibitor apocynin did not further alleviate HAPH, the mitochondrial antioxidant MitoQ did. Furthermore, in PASMCs, I3 attenuated hypoxia-induced PASMCs proliferation and reversed the activated HIF-1α/NOX4/PPAR-γ axis under hypoxia. Targeting this axis offset the protective effect of I3 on hypoxia-induced PASMCs proliferation. The present study is novel in revealing a new mechanism for preventing HAPH and provides insights into the optimization of intermittent short-duration reoxygenation.

DeepMPSF: A Deep Learning Network for Predicting General Protein Phosphorylation Sites Based on Multiple Protein Sequence Features.

Phosphorylation, as one of the most important post-translational modifications, plays a key role in various cellular physiological processes and disease occurrences. In recent years, computer technology has been gradually applied to the prediction of protein phosphorylation sites. However, most existing methods rely on simple protein sequence features that provide limited contextual information. To overcome this limitation, we propose DeepMPSF, a phosphorylation site prediction model based on multiple protein sequence features. There are two types of features: sequence semantic features, which comprise protein residue type information and relative position information within protein sequence, and protein background biophysical features, which include global semantic information containing more comprehensive protein background information obtained from pretrained models. To extract these features, DeepMPSF employs two separate subnetworks: the S71SFE module and the BBFE module, which automatically extract high-level semantic features. Our model incorporates a learning strategy for handling imbalanced datasets through ensemble learning during training and prediction. DeepMPSF is trained and evaluated on a well-established dataset of human proteins. Comparing the analysis with other benchmark methods reveals that DeepMPSF outperforms in predicting both S/T residues and Y residues. In particular, DeepMPSF showed excellent generalization performance in cross-species blind test performance, with an average improvement of 5.63%/5.72%, 22.28%/25.94%, 20.11%/17.49%, and 26.40%/28.33% for Mus musculus/Rattus norvegicus test sets in area under curves (AUCs) of ROC curve, AUC of the PR curve, F1-score, and MCC metrics, respectively. Furthermore, it also shows excellent performance in the latest updated case of natural proteins with functional phosphorylation sites. Through an ablation study and visual analysis, we uncover that the design of different feature modules significantly contributes to the accurate classification of DeepMPSF, which provides valuable insights for predicting phosphorylation sites and offers effective support for future downstream research.

A forensic population database of autosomal STR and X-STR markers in the Qiang ethnic minority of China.

The Qiang ethnic group is one of the oldest ethnic groups in China and is the most active ethnic group among all the populations along the Tibetan-Yi corridor. They have had a profound impact nationally and internationally. The paternal and maternal genetic feature of the Qiang ethnic group has been revealed, leaving the question of the genetic characteristics from autosomes and X chromosome not answered. The aim of this study was to explore the potential of 36 A-STR (Microreader™ 36A ID System) and 19 X-STR (Microreader™ 19X System) for application in the Qiang population and to elucidate their genetic diversity in southwest China. The cumulative probability of exclusion (CPE) for autosomal STRs is 1-1.3814 × 10-15 and the mean paternity exclusion chance (MEC) for X-STRs is 1-1.7323 × 10-6. Forensic parameters suggest that the STRs analyzed here are well-suited for forensic applications. The results of phylogenetic, interpopulation differentiation, and principal coordinates analysis (PCoA) indicate that the Qiang people have extensive connections with ethnic minorities in China, supporting the view that the Qiang people are the oldest group in the entire Sino-Tibetan language family. The Qiang appeared genetically more associated with most ethnic groups in China, especially the Han. The calculation of random matching probability (RMP) was improved by Fst correction of allele frequencies to make RMP more accurate and reasonable. This study can fill in the gaps in the Qiang STR reference database, providing valuable frequency data for forensic applications and evidence for the Qiang's genetic pattern as an important ancestral position in the Sino-Tibetan populations.

Expanding annotation of chemical compounds in hawthorn fruits and their variations in thermal processing using integrated mass spectral similarity networking.

Chemical structural characterization of chemical compounds from hawthorn fruits and its thermal processed products was carried out in present study. By linking Global Natural Products Social (GNPS) Molecular Networking and MolNetEnhancer workflow, seventy-four chemical compounds in hawthorn fruits and its thermal processed products were tentatively identified. Three quercetagetin derivatives (quercetagetin-3-O-glucoside, quercetagetin-di-glucoside and its isomer), five quercetin or kaempferol derivatives (quercetin-acetylapiosyl-hexoside, quercetin-3-O-(6″-malonyl-hexoside), quercetin-3-O-(6″-malonyl-hexoside)-(1 â†’ 2)-O-hexoside, quercetin-3-O-(6″-malonyl-hexoside)-(1 â†’ 2)-O-deoxyhexoside, kaempferol-3-O-(6″-malonyl-hexoside)), six procyanidins including four (E)C-ethyl-procyanidins and two A-type procyanidins digallate, as well as 13 triterpenoids including ursolic aldehyde, triterpenoid glycosides, and triterpene acids were reported for the first time in hawthorn fruits. In addition, triterpenoids exhibited considerable thermal stability, while all of flavonoid glycosides, proanthocyanidins and 10 in 13 organic acids showed dramatic decrease after thermal processing.

TransRNAm: Identifying Twelve Types of RNA Modifications by an Interpretable Multi-Label Deep Learning Model Based on Transformer.

Accurate identification of RNA modification sites is of great significance in understanding the functions and regulatory mechanisms of RNAs. Recent advances have shown great promise in applying computational methods based on deep learning for accurate prediction of RNA modifications. However, those methods generally predicted only a single type of RNA modification. In addition, such methods suffered from the scarcity of the interpretability for their predicted results. In this work, a new Transformer-based deep learning method was proposed to predict multiple RNA modifications simultaneously, referred to as TransRNAm. More specifically, TransRNAm employs Transformer to extract contextual feature and convolutional neural networks to further learn high-latent feature representations of RNA sequences relevant for RNA modifications. Importantly, by integrating the self-attention mechanism in Transformer with convolutional neural network, TransRNAm is capable of not only capturing the critical nucleotide sites that contribute significantly to RNA modification prediction, but also revealing the underlying association among different types of RNA modifications. Consequently, this work provided an accurate and interpretable predictor for multiple RNA modification prediction, which may contribute to uncovering the sequence-based forming mechanism of RNA modification sites.

MCnebula: Critical Chemical Classes for the Classification and Boost Identification by Visualization for Untargeted LC-MS/MS Data Analysis.

Untargeted mass spectrometry is a robust tool for biology, but it usually requires a large amount of time on data analysis, especially for system biology. A framework called Multiple-Chemical nebula (MCnebula) was developed herein to facilitate the LC-MS data analysis process by focusing on critical chemical classes and visualization in multiple dimensions. This framework consists of three vital steps as follows: (1) abundance-based classes (ABC) selection algorithm, (2) critical chemical classes to classify "features" (corresponding to compounds), and (3) visualization as multiple Child-Nebulae (network graph) with annotation, chemical classification, and structure. Notably, MCnebula can be used to explore the classification and structural characteristic of unknown compounds beyond the limit of the spectral library. Moreover, it is intuitive and convenient for pathway analysis and biomarker discovery because of its function of ABC selection and visualization. MCnebula was implemented in the R language. A series of tools in R packages were provided to facilitate downstream analysis in an MCnebula-featured way, including feature selection, homology tracing of top features, pathway enrichment analysis, heat map clustering analysis, spectral visualization analysis, chemical information query, and output analysis reports. The broad utility of MCnebula was illustrated by a human-derived serum data set for metabolomics analysis. The results indicated that "Acyl carnitines" were screened out by tracing structural classes of biomarkers, which was consistent with the reference. A plant-derived data set was investigated to achieve a rapid annotation and discovery of compounds in E. ulmoides.

SFTSV infection is associated with transient overproliferation of monoclonal lambda-type plasma cells.

The impairment of antibody-mediated immunity is a major factor associated with fatal cases of severe fever with thrombocytopenia syndrome (SFTS). By collating the clinical diagnosis reports of 30 SFTS cases, we discovered the overproliferation of monoclonal plasma cells (MCP cells, CD38+cLambda+cKappa-) in bone marrow, which has only been reported previously in multiple myeloma. The ratio of CD38+cLambda+ versus CD38+cKappa+ in SFTS cases with MCP cells was significantly higher than that in normal cases. MCP cells presented transient expression in the bone marrow, which was distinctly different from multiple myeloma. Moreover, the SFTS patients with MCP cells had higher clinical severity. Further, the overproliferation of MCP cells was also observed in SFTS virus (SFTSV)-infected mice with lethal infectious doses. Together, SFTSV infection induces transient overproliferation of monoclonal lambda-type plasma cells, which have important implications for the study of SFTSV pathogenesis, prognosis, and the rational development of therapeutics.

DGCddG: Deep Graph Convolution for Predicting Protein-Protein Binding Affinity Changes Upon Mutations.

Effectively and accurately predicting the effects of interactions between proteins after amino acid mutations is a key issue for understanding the mechanism of protein function and drug design. In this study, we present a deep graph convolution (DGC) network-based framework, DGCddG, to predict the changes of protein-protein binding affinity after mutation. DGCddG incorporates multi-layer graph convolution to extract a deep, contextualized representation for each residue of the protein complex structure. The mined channels of the mutation sites by DGC is then fitted to the binding affinity with a multi-layer perceptron. Experiments with results on multiple datasets show that our model can achieve relatively good performance for both single and multi-point mutations. For blind tests on datasets related to angiotensin-converting enzyme 2 binding with the SARS-CoV-2 virus, our method shows better results in predicting ACE2 changes, may help in finding favorable antibodies. Code and data availability: https://github.com/lennylv/DGCddG.

Simultaneous Prediction of Interaction Sites on the Protein and Peptide Sides of Complexes through Multilayer Graph Convolutional Networks.

Identifying the binding residues of protein-peptide complexes is essential for understanding protein function mechanisms and exploring drug discovery. Recently, many computational methods have been developed to predict the interaction sites of either protein or peptide. However, to our knowledge, no prediction method can simultaneously identify the interaction sites on both the protein and peptide sides. Here, we propose a deep graph convolutional network (GCN)-based method called GraphPPepIS to predict the interaction sites of protein-peptide complexes using protein and peptide structural information. We also propose a companion method, SeqPPepIS, for assisting with the lack of structural information and the flexibility of peptides. SepPPepIS replaces the peptide structural features in GraphPPepIS by learning features from peptide sequences. We performed a comprehensive evaluation of the benchmark data sets, and the results show that our two methods outperform state-of-the-art methods on the accurate interaction sites of both protein and peptide sides. We show that our methods can help improve protein-peptide docking. For docking data sets, our methods maintain robust performance in identifying binding sites, thereby enhancing the prediction of peptide binding poses. Finally, we visualized the analysis of protein and peptide graph embedding to demonstrate the learning ability of graph convolution in predicting interaction sites, which was mainly obtained through the shared parameters of a protein graph and peptide graph.

The Role of eIF5A1 in LPS-Induced Neuronal Remodeling of the Nucleus Accumbens in the Depression.

BACKGROUND: The pathogenesis of depression is complex, with the brain's reward system likely to play an important role. The nucleus accumbens (NAc) is a key region in the brain that integrates reward signals. Lipopolysaccharides (LPS) can induce depressive-like behaviors and enhance neuroplasticity in NAc, but the underlying mechanism is still unknown. We previously found that eukaryotic translation initiation factor A1 (eIF5A1) acts as a ribosome-binding protein to regulate protein translation and to promote neuroplasticity. METHODS: In the present study, LPS was administered intraperitoneally to rats and the expression and cellular location of eIF5A1 was then investigated by RT-PCR, Western blotting and immunofluorescence. Subsequently, a neuron-specific lentivirus was used to regulate eIF5A1 expression in vivo and in vitro. Neuroplasticity was then examined by Golgi staining and by measurement of neuronal processes. Finally, proteomic analysis was used to identify proteins regulated by eIF5A1. RESULTS: The results showed that eIF5A1 expression was significantly increased in the NAc neurons of LPS rats. Following the knockdown of eIF5A1 in NAc neurons, the LPS-induced increases in neuronal arbors and spine density were significantly attenuated. Depression-like behaviors were also reduced. Neurite outgrowth of NAc neurons in vitro also increased or decreased in parallel with the increase or decrease in eIF5A1 expression, respectively. The proteomic results showed that eIF5A1 regulates the expression of many neuroplasticity-related proteins in neurons. CONCLUSIONS: These results confirm that eIF5A1 is involved in LPS-induced depression-like behavior by increasing neuroplasticity in the NAc. Our study also suggests the brain's reward system may play an important role in the pathogenesis of depression.

Nanodrug rescues liver fibrosis via synergistic therapy with H2O2 depletion and Saikosaponin b1 sustained release.

Hypoxia and hydrogen peroxide (H2O2) accumulation form the profibrogenic liver environment, which involves fibrogenesis and chronic stimulation of hepatic stellate cells (HSCs). Catalase (CAT) is the major antioxidant enzyme that catalyzes H2O2 into oxygen and water, which loses its activity in different liver diseases, especially in liver fibrosis. Clinical specimens of cirrhosis patients and liver fibrotic mice are collected in this work, and results show that CAT decrease is closely correlated with hypoxia-induced transforminmg growth factor ß1 (TGF-ß1). A multifunctional nanosystem combining CAT-like MnO2 and anti-fibrosis Saikosaponin b1 (Ssb1) is subsequently constructed for antifibrotic therapy. MnO2 catalyzes the accumulated H2O2 into oxygen, thereby ameliorating the hypoxic and oxidative stress to prevent activation of HSCs, and assists to enhance the antifibrotic pharmaceutical effect of Ssb1. This work suggests that TGF-ß1 is responsible for the diminished CAT in liver fibrosis, and our designed MnO2@PLGA/Ssb1 nanosystem displays enhanced antifibrotic efficiency through removing excess H2O2 and hypoxic stress, which may be a promising therapeutic approach for liver fibrosis treatment.

Hypoxia Inhibits Cell Cycle Progression and Cell Proliferation in Brain Microvascular Endothelial Cells via the miR-212-3p/MCM2 Axis.

Hypoxia impairs blood-brain barrier (BBB) structure and function, causing pathophysiological changes in the context of stroke and high-altitude brain edema. Brain microvascular endothelial cells (BMECs) are major structural and functional elements of the BBB, and their exact role in hypoxia remains unknown. Here, we first deciphered the molecular events that occur in BMECs under 24 h hypoxia by whole-transcriptome sequencing assay. We found that hypoxia inhibited BMEC cell cycle progression and proliferation and downregulated minichromosome maintenance complex component 2 (Mcm2) expression. Mcm2 overexpression attenuated the inhibition of cell cycle progression and proliferation caused by hypoxia. Then, we predicted the upstream miRNAs of MCM2 through TargetScan and miRanDa and selected miR-212-3p, whose expression was significantly increased under hypoxia. Moreover, the miR-212-3p inhibitor attenuated the inhibition of cell cycle progression and cell proliferation caused by hypoxia by regulating MCM2. Taken together, these results suggest that the miR-212-3p/MCM2 axis plays an important role in BMECs under hypoxia and provide a potential target for the treatment of BBB disorder-related cerebrovascular disease.

CAPLA: improved prediction of protein-ligand binding affinity by a deep learning approach based on a cross-attention mechanism.

MOTIVATION: Accurate and rapid prediction of protein-ligand binding affinity is a great challenge currently encountered in drug discovery. Recent advances have manifested a promising alternative in applying deep learning-based computational approaches for accurately quantifying binding affinity. The structure complementarity between protein-binding pocket and ligand has a great effect on the binding strength between a protein and a ligand, but most of existing deep learning approaches usually extracted the features of pocket and ligand by these two detached modules. RESULTS: In this work, a new deep learning approach based on the cross-attention mechanism named CAPLA was developed for improved prediction of protein-ligand binding affinity by learning features from sequence-level information of both protein and ligand. Specifically, CAPLA employs the cross-attention mechanism to capture the mutual effect of protein-binding pocket and ligand. We evaluated the performance of our proposed CAPLA on comprehensive benchmarking experiments on binding affinity prediction, demonstrating the superior performance of CAPLA over state-of-the-art baseline approaches. Moreover, we provided the interpretability for CAPLA to uncover critical functional residues that contribute most to the binding affinity through the analysis of the attention scores generated by the cross-attention mechanism. Consequently, these results indicate that CAPLA is an effective approach for binding affinity prediction and may contribute to useful help for further consequent applications. AVAILABILITY AND IMPLEMENTATION: The source code of the method along with trained models is freely available at https://github.com/lennylv/CAPLA. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

ctP2ISP: Protein-Protein Interaction Sites Prediction Using Convolution and Transformer With Data Augmentation.

Protein-protein interactions are the basis of many cellular biological processes, such as cellular organization, signal transduction, and immune response. Identifying protein-protein interaction sites is essential for understanding the mechanisms of various biological processes, disease development, and drug design. However, it remains a challenging task to make accurate predictions, as the small amount of training data and severe imbalanced classification reduce the performance of computational methods. We design a deep learning method named ctP2ISP to improve the prediction of protein-protein interaction sites. ctP2ISP employs Convolution and Transformer to extract information and enhance information perception so that semantic features can be mined to identify protein-protein interaction sites. A weighting loss function with different sample weights is designed to suppress the preference of the model toward multi-category prediction. To efficiently reuse the information in the training set, a preprocessing of data augmentation with an improved sample-oriented sampling strategy is applied. The trained ctP2ISP was evaluated against current state-of-the-art methods on six public datasets. The results show that ctP2ISP outperforms all other competing methods on the balance metrics: F1, MCC, and AUPRC. In particular, our prediction on open tests related to viruses may also be consistent with biological insights. The source code and data can be obtained from https://github.com/lennylv/ctP2ISP.