MGNDTI: A Drug-Target Interaction Prediction Framework Based on Multimodal Representation Learning and the Gating Mechanism

Journal of Chemical Information and Modeling, Год журнала: 2024, Номер 64(16), С. 6684 - 6698

Опубликована: Авг. 13, 2024

Drug-Target Interaction (DTI) prediction facilitates acceleration of drug discovery and promotes repositioning. Most existing deep learning-based DTI methods can better extract discriminative features for drugs proteins, but they rarely consider multimodal drugs. Moreover, learning the interaction representations between targets needs further exploration. Here, we proposed a simple M ulti-modal G ating N etwork prediction, MGNDTI, based on representation gating mechanism. MGNDTI first learns sequence using different retentive networks. Next, it extracts molecular graph through convolutional network. Subsequently, devises network to obtain joint targets. Finally, builds fully connected computing probability. was benchmarked against seven state-of-the-art models (CPI-GNN, TransformerCPI, MolTrans, BACPI, CPGL, GIFDTI, FOTF-CPI) four data sets (i.e., Human, C. elegans, BioSNAP, BindingDB) under experimental settings. Through evaluation with AUROC, AUPRC, accuracy, F1 score, MCC, significantly outperformed above methods. is powerful tool showcasing its superior robustness generalization ability diverse It freely available at https://github.com/plhhnu/MGNDTI.

Язык: Английский

---

SCPLPA: An miRNA–disease association prediction model based on spatial consistency projection and label propagation algorithm

Journal of Cellular and Molecular Medicine, Год журнала: 2024, Номер 28(9)

Опубликована: Май 1, 2024

Abstract Identifying the association between miRNA and diseases is helpful for disease prevention, diagnosis treatment. It of great significance to use computational methods predict potential human associations. Considering shortcomings existing methods, such as low prediction accuracy weak generalization, we propose a new method called SCPLPA miRNA–disease First, heterogeneous similarity network was constructed using semantic Gaussian interaction spectrum kernel network, while functional network. Then, estimated scores were evaluated by integrating outcomes obtained implementing label propagation algorithms in Finally, spatial consistency projection algorithm used extract features unverified associations diseases. compared with four classical (MDHGI, NSEMDA, RFMDA SNMFMDA), results multiple evaluation metrics showed that exhibited most outstanding predictive performance. Case studies have shown can effectively identify miRNAs associated colon neoplasms kidney neoplasms. In summary, our proposed easy implement associations, making it reliable auxiliary tool biomedical research.

Язык: Английский

---

Unveiling patterns in spatial transcriptomics data: a novel approach utilizing graph attention autoencoder and multiscale deep subspace clustering network

GigaScience, Год журнала: 2025, Номер 14

Опубликована: Янв. 1, 2025

Abstract Background The accurate deciphering of spatial domains, along with the identification differentially expressed genes and inference cellular trajectory based on transcriptomic (ST) data, holds significant potential for enhancing our understanding tissue organization biological functions. However, most clustering methods can neither decipher complex structures in ST data nor entirely employ features embedded different layers. Results This article introduces STMSGAL, a novel framework analyzing by incorporating graph attention autoencoder multiscale deep subspace clustering. First, STMSGAL constructs ctaSNN, cell type–aware shared nearest neighbor graph, using Louvian exclusively gene expression profiles. Subsequently, it integrates profiles ctaSNN to generate spot latent representations Lastly, implements clustering, differential analysis, inference, providing comprehensive capabilities thorough exploration interpretation. was evaluated against 7 methods, including SCANPY, SEDR, CCST, DeepST, GraphST, STAGATE, SiGra, four 10x Genomics Visium datasets, 1 mouse visual cortex STARmap dataset, 2 Stereo-seq embryo datasets. comparison showcased STMSGAL’s remarkable performance across Davies–Bouldin, Calinski–Harabasz, S_Dbw, ARI values. significantly enhanced layer resolutions accurately delineated domains breast cancer tissues, adult brain (FFPE), embryos. Conclusions serve as an essential tool bridging analysis disease pathology, offering valuable insights researchers field.

Язык: Английский

---

DO-GMA: An End-to-End Drug–Target Interaction Identification Framework with a Depthwise Overparameterized Convolutional Network and the Gated Multihead Attention Mechanism

Journal of Chemical Information and Modeling, Год журнала: 2025, Номер unknown

Опубликована: Янв. 28, 2025

Identification of potential drug-target interactions (DTIs) is a crucial step in drug discovery and repurposing. Although deep learning effectively deciphers DTIs, most learning-based methods represent features from only single perspective. Moreover, the fusion method protein needs further refinement. To address above two problems, this study, we develop novel end-to-end framework named DO-GMA for DTI identification by incorporating Depthwise Overparameterized convolutional neural network Gated Multihead Attention mechanism with shared-learned queries bilinear model concatenation. first designs depthwise overparameterized to learn representations their SMILES strings amino acid sequences. Next, it extracts 2D molecular graphs through graph network. Subsequently, fuses combining gated attention multihead Finally, takes fused as inputs builds multilayer perceptron classify unlabeled pairs (DTPs). was benchmarked against six newest prediction (CPI-GNN, BACPI, CPGL, DrugBAN, BINDTI, FOTF-CPI) under four different experimental settings on data sets (i.e., DrugBank, BioSNAP, C.elegans, BindingDB). The results show that significantly outperformed based AUC, AUPR, accuracy, F1-score, MCC. An ablation robust statistical analysis, sensitivity analysis parameters, visualization features, computational cost case validated powerful performance DO-GMA. In addition, predicted drug-protein DB00568 P06276, DB09118 Q9UQD0) could be interacting. freely available at https://github.com/plhhnu/DO-GMA.

Язык: Английский

---

Machine Learning–Based Multiparameter Performance Predictions of Ultra‐Dense WDM FSO–FTTx Systems Under Diverse Weather Conditions

International Journal of Optics, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Free‐space optical (FSO) communication is vital for modern wireless systems due to its high data rates, energy efficiency, secure transmission, and cost‐effectiveness. However, weather‐induced attenuation, turbulence, pointing errors affect performance. Recent advancements leverage machine learning (ML) predict enhance system reliability under adverse conditions, improving channel awareness quality of service (QoS). This study has applied modulation schemes, including on‐off keying with nonreturn‐to‐zero (OOK–NRZ), quadrature phase shift (QPSK), polarization division multiplexed QPSK (PDM–QPSK), (PolSK) an ultra‐dense wavelength (WDM) FSO‐fiber‐to‐the‐x modulate the across clear, fog, rain, hazy conditions. Furthermore, multiparameter performance predictions diverse weather conditions have been assessed using ML algorithms such as extreme (ELM), support vector (SVM), gradient boosting (GB). Mean squared error (MSE) coefficient determination ( R 2 ) statistical measures employed measure robustness ML. Modulation atmospheric input powers, FSO link length are fed features. At same time, primary modeling targets signal‐to‐noise ratio (OSNR), factor (QF), bit rate (BER), rate, received power. The simulation results demonstrated that GB achieved model’s best reduced MSE values OSNR, QF, BER, power 2.0203e − 03, 9.3905e 04, 3.5214e 12, 1.2527e respectively. Moreover, exceptional ‐squared 0.9997, 0.9998, 0.9987, 1,

Язык: Английский

---

SEnSCA: Identifying possible ligand‐receptor interactions and its application in cell–cell communication inference

Journal of Cellular and Molecular Medicine, Год журнала: 2024, Номер 28(9)

Опубликована: Май 1, 2024

Abstract Multicellular organisms have dense affinity with the coordination of cellular activities, which severely depend on communication across diverse cell types. Cell–cell (CCC) is often mediated via ligand‐receptor interactions (LRIs). Existing CCC inference methods are limited to known LRIs. To address this problem, we developed a comprehensive analysis tool SEnSCA by integrating single RNA sequencing and proteome data. mainly contains potential LRI acquisition strength evaluation. For acquiring LRIs, it first extracts features reduces feature dimension, subsequently constructs negative samples through K‐means clustering, finally acquires LRIs based Stacking ensemble comprising support vector machine, 1D‐convolutional neural networks multi‐head attention mechanism. During evaluation, conducts filtering then infers combining three‐point estimation approach computed better precision, recall, accuracy, F1 score, AUC AUPR under most conditions when predicting possible illustrate inferred network, provided three visualization options: heatmap, bubble diagram network diagram. Its application human melanoma tissue demonstrated its reliability in detection. In summary, offers useful freely available at https://github.com/plhhnu/SEnSCA .

Язык: Английский

---

MRDPDA: A multi‐Laplacian regularized deepFM model for predicting piRNA‐disease associations

Journal of Cellular and Molecular Medicine, Год журнала: 2024, Номер 28(17)

Опубликована: Сен. 1, 2024

Abstract PIWI‐interacting RNAs (piRNAs) are a typical class of small non‐coding RNAs, which essential for gene regulation, genome stability and so on. Accumulating studies have revealed that piRNAs significant potential as biomarkers therapeutic targets variety diseases. However current computational methods face the challenge in effectively capturing piRNA‐disease associations (PDAs) from limited data. In this study, we propose novel method, MRDPDA, predicting PDAs based on data multiple sources. Specifically, MRDPDA integrates deep factorization machine (deepFM) model with regularizations derived yet datasets, utilizing separate Laplacians instead simple average similarity network. Moreover, unified objective function to combine embedding loss about similarities is proposed ensure suitable prediction task. addition, balanced benchmark dataset piRPheno constructed autoencoder applied creating reliable negative set unlabeled dataset. Compared three latest methods, achieves best performance pirpheno terms five‐fold cross validation test independent set, case further demonstrate effectiveness MRDPDA.

Язык: Английский

---

StereoSiTE: a framework to spatially and quantitatively profile the cellular neighborhood organized iTME

GigaScience, Год журнала: 2024, Номер 13

Опубликована: Янв. 1, 2024

Abstract Background Spatial transcriptome (ST) technologies are emerging as powerful tools for studying tumor biology. However, existing analyzing ST data limited, they mainly rely on algorithms developed single-cell RNA sequencing and do not fully utilize the spatial information. While some have been data, often designed specific tasks, lacking a comprehensive analytical framework leveraging Results In this study, we present StereoSiTE, an that combines open-source bioinformatics with custom to accurately infer functional cell interaction intensity (SCII) within cellular neighborhood (CN) of interest. We applied StereoSiTE decode datasets from xenograft models found CN efficiently distinguished different contexts, while SCII analysis provided more precise insights into intercellular interactions by incorporating By applying multiple samples, successfully identified region dominated neutrophils, suggesting their potential role in remodeling immune microenvironment (iTME) after treatment. Moreover, revealed neutrophil-mediated communication, supported pathway enrichment, transcription factor regulon activities, protein–protein interactions. Conclusions represents promising unraveling mechanisms underlying treatment response iTME CN-based tissue domain identification SCII-inferred The software is be scalable, modular, user-friendly, making it accessible wide range researchers.

Язык: Английский

---

DP-site: A dual deep learning-based method for protein-peptide interaction site prediction

Methods, Год журнала: 2024, Номер 229, С. 17 - 29

Опубликована: Июнь 12, 2024

Язык: Английский

---

STPDA: Leveraging spatial-temporal patterns for downstream analysis in spatial transcriptomic data

Computational Biology and Chemistry, Год журнала: 2024, Номер 112, С. 108127 - 108127

Опубликована: Июнь 11, 2024

Язык: Английский

---