The prediction of drug–target interaction (DTI) is crucial to drug discovery. Although the interactions between the drug and target can be accurately verified by traditional biochemical experiments, the determination of DTI through biochemical experiments is a time-consuming, laborious, and expensive process. Therefore, we propose a learning-based framework named BG-DTI for drug–target interaction prediction. Our model combines two main approaches based on biological features and heterogeneous networks to identify interactions between drugs and targets. First, we extract original features from the sequence to encode each drug and target. Later, we further consider the relationships among various biological entities by constructing drug–drug similarity networks and target–target similarity networks. Furthermore, a graph convolutional network and a graph attention network in the graph representation learning module help us learn the features representation of drugs and targets. After obtaining the features from graph representation learning modules, these features are combined into fusion descriptors for drug–target pairs. Finally, we send the fusion descriptors and labels to a random forest classifier for predicting DTI. The evaluation results show that BG-DTI achieves an average AUC of 0.938 and an average AUPR of 0.930, which is better than those of five existing state-of-the-art methods. We believe that BG-DTI can facilitate the development of drug discovery or drug repurposing.

Background/Objectives: Structural magnetic resonance imaging (MRI) and 18-fluoro-deoxy-glucose positron emission tomography (PET) reveal the structural and functional information of the brain from different dimensions, demonstrating considerable clinical and practical value in the computer-aided diagnosis of Alzheimer’s disease (AD). However, the structure and semantics of different modal data are different, and the distribution between different datasets is prone to the problem of domain shift. Most of the existing methods start from the single-modal data and assume that different datasets meet the same distribution, but they fail to fully consider the complementary information between the multi-modal data and fail to effectively solve the problem of domain distribution difference. Methods: In this study, we propose a multi-modal deep domain adaptation (MM-DDA) model that integrates MRI and PET modal data, which aims to maximize the utilization of the complementarity of the multi-modal data and narrow the differences in domain distribution to boost the accuracy of AD classification. Specifically, MM-DDA comprises three primary modules: (1) the feature encoding module, which employs convolutional neural networks (CNNs) to capture detailed and abstract feature representations from MRI and PET images; (2) the multi-head attention feature fusion module, which is used to fuse MRI and PET features, that is, to capture rich semantic information between modes from multiple angles by dynamically adjusting weights, so as to achieve more flexible and efficient feature fusion; and (3) the domain transfer module, which reduces the distributional discrepancies between the source and target domains by employing adversarial learning training. Results: We selected 639 subjects from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and considered two transfer learning settings. In ADNI1→ADNI2, the accuracies of the four experimental groups, AD vs. CN, pMCI vs. sMCI, AD vs. MCI, and MCI vs. CN, reached 92.40%, 81.81%, 81.13%, and 85.45%, respectively. In ADNI2→ADNI1, the accuracies of the four experimental groups, AD vs. CN, pMCI vs. sMCI, AD vs. MCI, and MCI vs. CN, reached 94.73%, 81.48%, 85.48%, and 81.69%, respectively. Conclusions: MM-DDA is compared with other deep learning methods on two kinds of transfer learning, and the performance comparison results confirmed the superiority of the proposed method in AD prediction tasks.

The multiple types of high throughput genomics data create a potential opportunity to identify driver pattern in ovarian cancer, which will acquire some novel and clinical biomarkers for appropriate diagnosis and treatment to cancer patients. However, it is a great challenging work to integrate omics data, including somatic mutations, Copy Number Variations (CNVs) and gene expression profiles, to distinguish interactions and regulations which are hidden in drug response dataset of ovarian cancer. To distinguish the candidate driver genes and the corresponding driving pattern for resistant and sensitive tumor from the heterogeneous data, we combined gene co-expression modules and mutation modulators and proposed the identification driver patterns method. Firstly, co-expression network analysis is applied to explore gene modules for gene expression profiles via weighted correlation network analysis (WGCNA). Secondly, mutation matrix is generated by integrating the CNVs and somatic mutations, and a mutation network is constructed from this mutation matrix. The candidate modulators are selected from the significant genes by clustering the vertex of the mutation network. At last, regression tree model is utilized for module networks learning in which the achieved gene modules and candidate modulators are trained for the driving pattern identification and modulator regulatory exploring. Many of the candidate modulators identified are known to be involved in biological meaningful processes associated with ovarian cancer, which can be regard as potential driver genes, such as CCL11, CCL16, CCL18, CCL23, CCL8, CCL5, APOB, BRCA1, SLC18A1, FGF22, GADD45B, GNA15, GNA11 and so on, which can help to facilitate the discovery of biomarkers, molecular diagnostics, and drug discovery.

: The prediction of drug-target affinity (DTA) plays an important role in the development of drugs and the discovery of potential drug targets. In recent years, computer-assisted DTA prediction has become an important method in this field. In this work, we propose a multi-modal deep learning framework for drug-target binding affinity and binding region prediction, namely MMD-DTA. The model can predict DTA while unsupervised learning of drug-target binding regions. The experimental results show that MMD-DTA performs better than the existing models on the main evaluation metrics. In addition, external validation results show that MMD-DTA improves the generalization ability of the model by integrating sequence information and structural information of drugs and targets, and the model trained on the benchmark dataset can be well generalized to independent virtual screening tasks. Visualization of drug-target binding region prediction shows the powerful interpretability of MMD-DTA, which has important implications for exploring the functional regions of drug molecules acting on proteins.

Introduction: In the early asymptomatic stages of cancer, the immune system initiates a targeted response against tumor-associated antigens. During this phase, the immune system specifically identifies tumor antigens and triggers the clonal expansion of tumor antigen-specific T cells, which recognize tumor antigen peptides presented by the major histocompatibility complex via the T-cell receptor (TCR) on their surface. Consequently, monitoring alterations in the TCR repertoire holds promise for evaluating an individual’s immune status for cancer detection.Methods: In this study, we introduced a deep learning framework named DeepCaTCR, designed to enhance the prediction of cancer-associated T-cell receptors. The framework employs a one-dimensional convolutional neural network with variable convolutional kernels, a bidirectional long short-term memory network, and a self-attention mechanism to facilitate feature extraction from amino acid fragments of varying lengths.Results: DeepCaTCR demonstrates superior performance in cancer-associated TCR recognition, achieving an area under the receiver operating characteristic curve (AUC) of 0.863 and an F1-score of 0.669, thereby outperforming prevailing deep learning models. Validation result indicates that DeepCaTCR effectively distinguishes between tumor-infiltrating lymphocytes (TILs) and healthy peripheral blood samples, achieving an AUC greater than 0.95. It also exhibits high sensitivity (62.5%) and specificity (over 98%) in peripheral blood testing for early-stage cancer patients. To further enhance detection efficacy, we introduced a variance-based repertoire scoring strategy to quantify the dynamic heterogeneity of TCR clonal amplification, resulting in an increased AUC of 0.967 for pan-cancer early screening.Discussion: This study introduces a novel tool for analyzing the tumor immune microenvironment, offering significant translational potential for early cancer diagnosis. Its key feature is a new scoring method based on variance, not the average method.