Recent advancements in Spatial Transcriptomics (ST) technologies have enabled researchers to investigate the relationships between cells while simultaneously considering their spatial locations within tissue. These technologies facilitate the integration of gene expression data with spatial information for clustering analysis. While many clustering methods have been developed, they typically rely on the dataset’s intrinsic features without incorporating domain knowledge, such as marker genes. We argue that incorporating marker gene information can enhance the learning of cell embedding and improve clustering outcomes. In this paper, we introduce MGGNN (Marker Gene-Guided Graph Neural Networks), a novel approach designed to enhance spatial transcriptomics clustering. Firstly, we train the model using a contrastive learning framework based on a Graph Neural Network (GNN). Subsequently, we fine-tune the model using a few spots labeled by the expression of marker genes. Simulation and experiments conducted on two real-world datasets demonstrate the superior performance of our model over state-of-the-art methods.

Autism spectrum disorder (ASD) is a neuropathic disease which is characterized by deficits in social interaction and communication. Therefore, the ASD patients have weak ability to express themselves or let others know about their thoughts. As society pays more attention to ASD patients, early intervention programs, behavioral therapy and technological assistance have emerged to help ASD patients improve their quality of lives. This paper aims to propose an improved object detection algorithm based on Faster R-CNN-MobileNetV3 to analyze the micro expressions of ASD patients. The data set includes 1358 face images of ASD patients built from 12 ASD movies with the method of Cinemetrics. Through the training and testing of the ASD data set with the improved model, the overall precision rate has reached 0.9 and mean Average Precision also has significant improvement. As a result, the improved Faster R-CNN-MobileNetV3 model achieves a good performance to recognize micro expressions and emotions of ASD patients.

Traditional Chinese medicine (TCM) not only embodies a wealth of medical knowledge and cultural value but also exerts a profound influence on the global health domain. However, due to the antiquity, complexity, and uniqueness of its content, traditional computer-aided diagnosis methods struggle to harness its potential fully. In recent years, with the rapid development of artificial intelligence technology, large language models (LLMs) have gradually been applied to the field of TCM due to their outstanding natural language processing capabilities. This paper investigates the current applications of LLMs in TCM, covering various aspects such as clinical assistance, education assessment, and knowledge mining. We first review the construction of datasets related to TCM, including corpus data from TCM literature, medical case records, and clinical data. Next, we explore the specific applications of LLMs in integrating TCM knowledge, supporting clinical decision-making, and assessing education, analyzing their performance in improving efficiency and optimizing resources. Finally, we discuss the challenges faced by LLMs in TCM applications, such as data scarcity, model adaptability, and standardization of evaluation, and provide an outlook on future research and development prospects.