Predicting accurately the mechanisms of drug-drug interaction (DDI) events is crucial in drug research and development. Existing methods used to predict these events are primarily based on deep learning and have achieved satisfactory results. However, they rarely consider the presence of redundant co-information between the multimodal data of a drug and the need for consistency in the predicted features of each drug modality. Herein, we propose a new method for drug interaction event prediction based on multimodal mutual orthogonal projection and intermodal consistency loss. Our method obtains the features of each modality through a multimodal mutual orthogonal projection module, which eliminates redundant common information with other modalities. In addition, we use the consistency loss between modalities and make the predicted features of each modality more similar. In comparative experiments, our proposed method achieves a prediction accuracy of 0.9500, and an area under the precision-recall (AUPR) curve is 0.9833 for known DDIs. This method outperforms existing methods. The results show that the proposed method is capable of accurately predicting DDIs.

Abstract Background Conventional immunohistochemistry technologies were limited by the inability to simultaneously detect multiple markers and the lack of identifying spatial relationships among cells, hindering understanding of the biological processes in cancer immunology. Methods Tissue slices of primary tumours from 553 IA∼IIIB non‐small cell lung cancer (NSCLC) cases were stained by multiplex immunofluorescence (mIF) assay for 10 markers, including CD4, CD38, CD20, FOXP3, CD66b, CD8, CD68, PD‐L1, CD133 and CD163, evaluating the amounts of 26 phenotypes of cells in tumour nest and tumour stroma. StarDist depth learning model was utilised to determine the spatial location of cells based on mIF graphs. Single‐cell RNA sequencing (scRNA‐seq) on four primary NSCLC cases was conducted to investigate the putative cell interaction networks. Results Spatial proximity among CD20+ B cells, CD4+ T cells and CD38+ T cells (r2 = 0.41) was observed, whereas the distribution of regulatory T cells was associated with decreased infiltration levels of CD20+ B cells and CD38+ T cells (r2 = −0.45). Univariate Cox analyses identified closer proximity between CD8+ T cells predicted longer disease‐free survival (DFS). In contrast, closer proximity between CD133+ cancer stem cells (CSCs), longer distances between CD4+ T cells and CD20+ B cells, CD4+ T cells and neutrophils, and CD20+ B cells and neutrophils were correlated with dismal DFS. Data from scRNA‐seq further showed that spatially adjacent N1‐like neutrophils could boost the proliferation and activation of T and B lymphocytes, whereas spatially neighbouring M2‐like macrophages showed negative effects. An immune‐related risk score (IRRS) system aggregating robust quantitative and spatial prognosticators showed that high‐IRRS patients had significantly worse DFS than low‐IRRS ones (HR 2.72, 95% CI 1.87–3.94, p < .001). Conclusions We developed a framework to analyse the cell interaction networks in tumour microenvironment, revealing the spatial architecture and intricate interplays between immune and tumour cells.

Current convolutional neural network-based ultrasound automatic classification models for prostate cancer often rely on extensive manual labeling. Although Self-supervised Learning (SSL) have shown promise in addressing this problem, those data that from medical scenarios contains intra-class similarity conflicts, so using loss calculations directly that include positive and negative sample pairs can mislead training. SSL method tends to focus on global consistency at the image level and does not consider the internal informative relationships of the feature map. To improve the efficiency of prostate cancer diagnosis, using SSL method to learn key diagnostic information in ultrasound images, we proposed a self-supervised dual-head attentional bootstrap learning network (SDABL), including Online-Net and Target-Net. Self-Position Attention Module (SPAM) and adaptive maximum channel attention module (CAAM) are inserted in both paths simultaneously. They captures position and inter-channel attention and of the original feature map with a small number of parameters, solve the information optimization problem of feature maps in SSL. In loss calculations, we discard the construction of negative sample pairs, and instead guide the network to learn the consistency of the location space and channel space by drawing closer to the embedding representation of positive samples continuously. We conducted numerous experiments on the prostate Transrectal ultrasound (TRUS) dataset, experiments show that our SDABL pre-training method has significant advantages over both mainstream contrast learning methods and other attention-based methods. Specifically, the SDABL pre-trained backbone achieves 80.46% accuracy on our TRUS dataset after fine-tuning.

The computer-aided diagnosis of prostate ultrasound images can aid in the detection and treatment of prostate cancer. However, the ultrasound images of the prostate sometimes come with serious speckle noise, low signal-to-noise ratio, and poor detection accuracy. To overcome this shortcoming, we proposed a deep learning model that integrates S-Mask R-CNN and Inception-v3 in the ultrasound image-aided diagnosis of prostate cancer in this paper. The improved S-Mask R-CNN was used to realize the accurate segmentation of prostate ultrasound images and generate candidate regions. The region of interest align algorithm was used to realize the pixel-level feature point positioning. The corresponding binary mask of prostate images was generated by the convolution network to segment the prostate region and the background. Then, the background information was shielded, and a data set of segmented ultrasound images of the prostate was constructed for the Inception-v3 network for lesion detection. A new network model was added to replace the original classification module, which is composed of forward and back propagation. Forward propagation mainly transfers the characteristics extracted from the convolution layer pooling layer below the pool_3 layer through the transfer learning strategy to the input layer and then calculates the loss value between the classified and label values to identify the ultrasound lesion of the prostate. The experimental results showed that the proposed method can accurately detect the ultrasound image of the prostate and segment prostate information at the pixel-level simultaneously. The proposed method has higher accuracy than that of the doctor’s manual diagnosis and other detection methods. Our simple and effective approach will serve as a solid baseline and help ease future research in the computer-aided diagnosis of prostate ultrasound images. Furthermore, this work will promote the development of prostate cancer ultrasound diagnostic technology.

Vehicle routing problem with simultaneous delivery and pickup and time windows (VRPSDPTW) is an important logistics problem in supply chain network optimization. However, at present, existing approaches are only applied to solve small-scale many-objective VRPSDPTW. For the large-scale problem, to obtain high-quality solutions (effectively) and solve the problem in a short running time (efficiently), we focus on two issues: 1) how to handle many-objective property and 2) how to tackle large-scale property of the problem. For the first issue, to enhance the convergence property, a novel local search algorithm utilizing seven neighborhood operators is proposed. Meanwhile, to maintain the diversity property, the concept of decomposition is adopted. The problem is first decomposed into multiple single-objective problems, then local search is utilized to optimize these subproblems. For the second issue, since optimizing a subproblem by local search is time consuming for large-scale problems, the key point is how to allocate the computational resources and balance the convergence and diversity. To handle this problem, a weight space partition strategy is introduced. Finally, a decomposition-based local search algorithm is proposed for many-objective VRPSDPTW. Experimental results on the large-scale instances provided by Chinese logistic company show the effectiveness of the proposed algorithm.

Glaucoma is a chronic and widespread eye disease threatening humans' irreversible vision loss. The cup-to-disc ratio (CDR), one of the most important measurements used for glaucoma screening and diagnosis, requires accurate segmentation of optic disc and cup from fundus images. However, most existing techniques fail to obtain satisfactory segmentation performance because a significant number of pixel-level annotated data are often unavailable during training. To cope with this limitation, in this paper, we propose an effective joint optic disc and cup segmentation method based on semi-supervised conditional Generative Adversarial Nets (GANs). Our architecture consists of a segmentation net, a generator and a discriminator, to learn a mapping between the fundus images and the corresponding segmentation maps. Additionally, we employ both labeled and unlabeled data to improve the segmentation performance. The extensive experiments show that our method achieves state-of-the-art optic disc and cup segmentation results on both ORIGA and REFUGE datasets.

Metaheuristics have been widely utilized for solving NP-hard optimization problems. However, these algorithms usually perform differently from one problem to another, i.e., one may be effective on a problem but performs badly on another problem. Therefore, it is difficult to choose the best algorithm in advance for a given problem. In contrast to selecting the best algorithm for a problem, selection hyper-heuristics aim at performing well on a set of problems (instances). This paper proposes a selection hyper-heuristic based algorithm for multi-objective optimization problems. In the proposed algorithm, multiple metaheuristics exhibiting different search behaviors are managed and controlled as low-level metaheuristics in an algorithm pool, and the most appropriate metaheuristic is selected by means of a performance indicator at each search stage. To assess the performance of the proposed algorithm, an implementation of the algorithm containing four metaheuristics is proposed and tested for solving multi-objective unconstrained binary quadratic programming problem. Experimental results on 50 benchmark instances show that the proposed algorithm can provide better overall performance than single metaheuristics, which demonstrates the effectiveness of the proposed algorithm.