Protein-ligand binding affinity prediction is critical for drug discovery, yet existing methods struggle to jointly model local atomic interactions and global contextual dependencies. To address this, we propose the Interpretable Dual-Branch Graph-Transformer framework for Protein-Ligand Affinity prediction (DBGT-PLA), a novel dual-branch architecture that integrates graph neural network (GNN) with a stability-enhanced Transformer equipped with learnable positional embeddings and a NaN-filtering mechanism that handles potential Not-a-Number (NaN) values arising from numerical instability or data preprocessing. We design a Gated Residual Learning (GRL) Fusion module that performs dimension-wise adaptive integration between local graph topology and global Transformer context. This mechanism enables multi-level feature coordination through a residual path, achieving biophysically consistent alignment between atomic-level interactions and global conformational dependencies. Furthermore, we introduce an edge-level Shapley attribution framework tailored to protein-ligand interaction graphs, quantifying contributions of chemical bonds (e.g., hydrophobic contacts) and non-covalent interactions. Experiments show DBGT-PLA reduces RMSE by 18.3% (from 1.522 to 1.244 on the Holdout Set 2019), outperforming state-of-the-art models. Crucially, our explainability module reveals that the ligand edges dominate affinity predictions, accounting for nearly 70%. This work not only advances predictive accuracy but also offers unprecedented, quantitative insights into interaction determinants, which can guide rational drug optimization.

Single-cell sequencing technology has enabled researchers to study cellular heterogeneity at the cell level. To facilitate the downstream analysis, clustering single-cell data into subgroups is essential. However, the high dimensionality, sparsity, and dropout events of the data make the clustering challenging. Currently, many deep learning methods have been proposed. Nevertheless, they either fail to fully utilize pairwise distances information between similar cells, or do not adequately capture their feature correlations. They cannot also effectively handle high-dimensional sparse data. Therefore, they are not suitable for high-fidelity clustering, leading to difficulties in analyzing the clear cell types required for downstream analysis. The proposed scSAMAC method integrates contrastive learning and negative binomial losses into a variational autoencoder, extracting features via contrastive unit similarity while preserving the intrinsic characteristics. This enhances the robustness and generalization during the clustering. In the contrastive learning, it constructs a mask module by adopting a negative sample generation method with gene feature saliency adjustment, which selects features more influential in the clustering phase and simulates data missing events. Additionally, it develops a novel loss, which consists of a soft k-means loss, a Wasserstein distance, and a contrastive loss. This fully utilizes data information and improves clustering performance. Furthermore, a multi-head attention mechanism module is applied to the latent variables at each layer of autoencoder to enhance feature correlation, integration, and information repair. Experimental results demonstrate that scSAMAC outperforms several state-of-the-art clustering methods.

The emergence of single-cell RNA-seq (scRNA-seq) technology makes it possible to capture their differences at the cellular level, which contributes to studying cell heterogeneity. By extracting, amplifying and sequencing the genome at the individual cell level, scRNA-seq can be used to identify unknown or rare cell types as well as genes differentially expressed in specific cell types under different conditions using clustering for downstream analysis of scRNA-seq. Many clustering algorithms have been developed with much progress. However, scRNA-seq often appears with characteristics of high dimensions, sparsity and even the case of dropout events', which make the performance of scRNA-seq data clustering unsatisfactory. To circumvent the problem, a new deep learning framework, termed variational graph attention auto-encoder (VGAAE), is constructed for scRNA-seq data clustering. In the proposed VGAAE, a multi-head attention mechanism is introduced to learn more robust low-dimensional representations for the original scRNA-seq data and then self-supervised learning is also recommended to refine the clusters, whose number can be automatically determined using Jaccard index. Experiments have been conducted on different datasets and results show that VGAAE outperforms some other state-of-the-art clustering methods.

Vehicular ad-hoc networks (VANETs) have been widely studied in intelligent transportation. Content delivery is an important topic that attracts many researchers. Due to vehicles that may have intermittent connections and uncertain routes, it is difficult to select an appropriate node. In this paper, we analyze the movement pattern of vehicles from real taxis’ trajectories and propose a framework for delivery prediction, which aims to select appropriate nodes. First, we propose the framework which consists of a contact clique model, a social clique model, and a prediction model based on Markov chains, to characterize the movement pattern of vehicles. Second, we capture dynamic movement patterns by dividing the time requirement into equal length slots and construct clique sequences. Based on the fact that the sociality of nodes has strong temporal correlations, we utilize the prediction model to derive future cliques and evaluate two kinds of delivery performance in the future. Finally, we design a content delivery algorithm with dynamic movement pattern (DMP) to select the appropriate node. In our experiment, DMP performs better than that of other methods in terms of overhead, average hops. Also, as the number of nodes increases, our algorithm keeps small fluctuations in node sociality.

Lymph-node metastasis is the most perilous cancer progressive state, where long non-coding RNA (lncRNA) has been confirmed to be an important genetic indicator in cancer prediction. However, lncRNA expression profile is often characterized of large features and small samples, it is urgent to establish an efficient judgment to deal with such high dimensional lncRNA data, which will aid in clinical targeted treatment. Thus, in this study, a local linear reconstruction guided distance metric learning is put forward to handle lncRNA data for determination of cancer lymph-node metastasis. In the original locally linear embedding (LLE) approach, any point can be approximately linearly reconstructed using its nearest neighborhood points, from which a novel distance metric can be learned by satisfying both nonnegative and sum-to-one constraints on the reconstruction weights. Taking the defined distance metric and lncRNA data supervised information into account, a local margin model will be deduced to find a low dimensional subspace for lncRNA signature extraction. At last, a classifier is constructed to predict cancer lymph-node metastasis, where the learned distance metric is also adopted. Several experiments on lncRNA data sets have been carried out, and experimental results show the performance of the proposed method by making comparisons with some other related dimensionality reduction methods and the classical classifier models.

BACKGROUND: In the process of designing drugs and proteins, it is crucial to recognize hot regions in protein-protein interactions. Each hot region of protein-protein interaction is composed of at least three hot spots, which play an important role in binding. However, it takes time and labor force to identify hot spots through biological experiments. If predictive models based on machine learning methods can be trained, the drug design process can be effectively accelerated. RESULTS: The results show that different machine learning algorithms perform similarly, as evaluating using the F-measure. The main differences between these methods are recall and precision. Since the key attribute of hot regions is that they are packed tightly, we used the cluster algorithm to predict hot regions. By combining Gaussian Naïve Bayes and DBSCAN, the F-measure of hot region prediction can reach 0.809. CONCLUSIONS: In this paper, different machine learning models such as Gaussian Naïve Bayes, SVM, Xgboost, Random Forest, and Artificial Neural Network are used to predict hot spots. The experiment results show that the combination of hot spot classification algorithm with higher recall rate and clustering algorithm with higher precision can effectively improve the accuracy of hot region prediction.

In the field of medical images, skin lesion segmentation in dermoscopic images is a challenging task due to the irregular and blurring edges of the lesion and the presence of various artifacts. With the successful application of generative antagonistic network (GAN), a new neural network for skin lesion segmentation is proposed. The encoder-decoder with Dense-Residual block is used in the segmentation network which enables the network to be trained more efficiently. A multi-scale objective loss function is introduced to utilize deep supervision. We combine Jaccard distance and End Point Error which can solve lesion-background imbalance problem in pixel-level classification for skin lesion segmentation and also alleviate the problem of boundary ambiguity. A joint loss function is finally used, which includes a multi-scale objective loss function, End Point Error and Jaccard distance content loss function. Experiment results show that our algorithm is superior to other state-of-the-art algorithms on the ISBI2017.

Optical coherence tomography (OCT) is a primary imaging technique for ophthalmic diagnosis, which has the advantages of high-resolution and non-invasive. Diabetes is a chronic disease which might increase the risk of blindness. Hence, it is important to monitor the morphology of the retinal layer and fluid accumulation for Diabetic macular edema (DME) patients. In this paper, we proposed a new semi-supervised fully convolutional deep learning approach for segmenting retinal layers and fluid region in retinal OCT B-scans. The proposed semi -supervised approach leverages unlabeled data through an adversarial learning strategy. The segmentation framework includes a segment network and a discriminate network, both two networks are u-net like fully convolutional architecture. The objective function of the segment network is a joint loss function including multi-class cross entropy loss, adversarial loss and semi-supervise loss. Experiment result on the duke DME dataset demonstrate the effectiveness of the proposed segmentation framework.

Dimensionality reduction is indispensable for high dimensional data classification. So in this paper, a novel supervised method is developed to reduce dimensions of the original data, which is named feature space distance metric learning (FSDML). Instead of distances between any two points, distances between any two feature spaces are involved in the proposed method. Besides feature space distances(FSD) metric, the inter-class data separablity and the intra-class data locality are all employed for graph embedding, by which a subspace will be explored for data discriminant analysis. The proposed FSDML are evaluated by some state-of-art methods such as linear discriminant analysis (LDA), unsupervised discriminant projection (UDP) and nearest feature space embedding (NFSE). Experiments on some benchmark data sets including AR and ORL face data have shown that the proposed method is effective and efficient.

Saliency computational model with active environment perception can substantially facilitate a wide range of applications. Conventional saliency computational models primarily rely on hand-crafted low level image features, such as color or contrast. However, they may face great challenges in low lighting scenario, due to the lack of well-defined feature to represent saliency information in low contrast images. In this paper, we propose a novel deep neural network framework embedded with covariance descriptor for salient object detection in low contrast images. Several low-level features are extracted to compute their mutual covariance, which is then trained via a 7-layers convolutional neural network (CNN). The saliency map can be generated by estimating the final saliency score of each region via the pre-trained CNN model. Extensive experiments have been conducted on six challenging datasets to evaluate the performance of the proposed model against ten state-of-the-art models.

In this paper, a novel supervised dimensionality reduction method based on LLE is put forward, which is titled locally linear representation Fisher criterion (LLRFC). In the proposed LLRFC, the class information of the original data has been fully considered, according to which an inter-class graph and an intra-class graph can be well modeled respectively. Meanwhile, the neighborhoods in the inter-class graph consist of samples with various labels and the neighborhoods in the intra-graph are just composed of points sampled from the same class. Then the least locally linear representation technique is introduced to optimize the reconstruction weights in both graphs. At last, the Fisher criterion with maximum inter-class scatter and minimum intra-class scatter is reasoned. Experiments on some benchmark face data sets have been conducted and the results validate the proposed method's performance.

In this paper, an efficient feature extraction method named as Constrained Maximum Variance Mapping (CMVM) is developed for dimensionality reduction. The proposed algorithm can be viewed as a linear approximation of multi-manifolds based learning approach, which takes the local geometry and manifold labels into account. After the local scatters have been characterized, the proposed method focuses on developing a linear transformation that can maximize the distances matrix between all the manifolds under the constraint of locality preserving. Then, YALE face database, ORL face database are all taken to examine the effectiveness and efficiency of the proposed method. Experimental results validate that the proposed approach is superior to other widely used feature extraction methods.

ISOMAP is a manifold learning based algorithm for dimensionality reduction, which is successfully applied to data visualization. However, there exists such limitation in classical ISOMAP that the algorithm is sensitive to noises, especially outliers. So in this paper an extended ISOMAP algorithm is put forward to solve the problem of sensitivity. The proposed algorithm follows the method of classical ISOMAP except that a preprocessing strategy is introduced to remove the noises and outliers. The likelihood of each point to be a noise or an outlier is quantified by carrying out weighted principal component analysis and box statistics method is adopted to distinguish clear points from noisy ones, then ISOMAP can be performed after de-noising. Experiments on noisy s-curve and noisy Swiss-roll data validate its efficiency for improving robustness.

It is a challenging task to diagnose tumor type precisely based on microarray data because the number of variables p (genes) is far larger than that of samples, n. Many independent component analysis (ICA) based models had been proposed to tackle the microarray data classification problem with great success. Although it was pointed out that different independent components (ICs) are of different biological significance, up to now, it is still far from well explored for the problem that how to select proper IC subsets to predict new samples best. We try to improve the performance of ICA based classification models by using proper IC subsets instead of all the ICs. A genetic algorithms (GA) based selection process is proposed in this paper, and the selected IC subset is evaluated by the leave-one-out cross validation (LOOCV) technique. The experimental results demonstrate that our GA based IC selection method can further improve the classification accuracy of the ICA based prediction models.