Predicting circRNA–Disease Associations through Multisource Domain-Aware Embeddings and Feature Projection Networks

Journal of Chemical Information and Modeling, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 19, 2025

Recent studies have highlighted the significant role of circular RNAs (circRNAs) in various diseases. Accurately predicting circRNA–disease associations is crucial for understanding their biological functions and disease mechanisms. This work introduces MNDCDA method, designed to address challenges posed by limited number known high cost experiments. integrates multiple data sources with neighborhood-aware embedding models deep feature projection networks predict potential pathways linking circRNAs Initially, comprehensive biometric are used construct four similarity networks, forming a diverse interaction framework. Next, model captures structural information about diseases, while learn high-order interactions nonlinear connections. Finally, bilinear decoder identifies novel between The achieved an AUC 0.9070 on constructed benchmark dataset. In case studies, 25 out 30 predicted pairs were validated through wet lab experiments published literature. These extensive experimental results demonstrate that robust computational tool associations, providing valuable insights helping reduce research costs.

Language: Английский

---

Accurate identification of snoRNA targets using variational graph autoencoder to advance the redevelopment of traditional medicines

Frontiers in Pharmacology, Journal Year: 2025, Volume and Issue: 15

Published: Jan. 6, 2025

Existing studies indicate that dysregulation or abnormal expression of small nucleolar RNA (snoRNA) is closely associated with various diseases, including lung cancer. Furthermore, these diseases often involve multiple targets, making the redevelopment traditional medicines highly promising. Accurate prediction potential snoRNA therapeutic targets essential for early disease intervention and medicines. Additionally, researchers have developed artificial intelligence (AI)-based methods to screen predict thereby advancing drug redevelopment. However, existing face challenges such as imbalanced datasets dominance high-degree nodes in graph neural networks (GNNs), which compromise accuracy node representations. To address challenges, we propose an AI model based on variational autoencoders (VGAEs) integrates decoupling Kolmogorov-Arnold Network (KAN) technologies. The reconstructs snoRNA-disease graphs by learning representations, accurately identifying targets. By similarity from degree, mitigates nodes, enhances scenarios like cancer, leverages KAN technology improve adaptability flexibility new data. Case revealed SNORA21 SNORD33 are abnormally expressed cancer patients strong candidates These findings validate proposed model's effectiveness supporting screening treatment, Data experimental archived in: https://github.com/shmildsj/data.

Language: Английский

---

Semi-Correlations for the Simulation of Dermal Toxicity

Toxics, Journal Year: 2025, Volume and Issue: 13(4), P. 235 - 235

Published: March 23, 2025

The skin is the primary pathway for harmful substances to enter body and a susceptible target organ, making compound-induced acute dermal toxicity significant health risk. In this work, possibility of modelling using so-called semi-correlations studied. Semi-correlations are specific case correlations, where one variable takes only two values. For example, 0 denotes absence activity (e.g., toxicity), 1 presence activity. described computational experiments can be carried out by interested readers freely available software CORAL.

Language: Английский

---

MKAN-MMI: empowering traditional medicine-microbe interaction prediction with masked graph autoencoders and KANs

Frontiers in Pharmacology, Journal Year: 2024, Volume and Issue: 15

Published: Oct. 22, 2024

The growing microbial resistance to traditional medicines necessitates in-depth analysis of medicine-microbe interactions (MMIs) develop new therapeutic strategies. Widely used artificial intelligence models are limited by sparse observational data and prevalent noise, leading over-reliance on specific for feature extraction reduced generalization ability. To address these limitations, we integrate Kolmogorov-Arnold Networks (KANs), independent subspaces, collaborative decoding techniques into the masked graph autoencoder (Mask GAE) framework, creating an innovative MMI prediction model with enhanced accuracy, generalization, interpretability. First, apply Bernoulli distribution randomly mask parts graph, advancing self-supervised training reducing noise impact. Additionally, subspace technique enables neural networks (GNNs) learn weights independently across different enhancing expression. Fusing multi-layer outputs GNNs effectively reduces information loss caused masking. Moreover, using KANs advanced nonlinear mapping enhances learnability interpretability weights, deepening understanding complex MMIs. These measures significantly our in tasks. We validated three public datasets results showing that outperformed existing models. relevant code publicly accessible at: https://github.com/zhuoninnin1992/MKAN-MMI.

Language: Английский

---

Graph pooling for graph-level representation learning: a survey

Artificial Intelligence Review, Journal Year: 2024, Volume and Issue: 58(2)

Published: Dec. 20, 2024

In graph-level representation learning tasks, graph neural networks have received much attention for their powerful feature capabilities. However, with the increasing scales of data, how to efficiently process and extract key information has become focus research. The pooling technique, as a step in networks, simplifies structure by merging nodes or subgraphs, which significantly improves computational efficiency extraction ability networks. Although various methods been proposed numerous scholars, there is still relative lack systematic summaries these works. this paper, we comprehensively sort out fundamentals recent progress techniques discuss its wide range application scenarios, well current challenges opportunities, point direction future Specifically, first provide detailed introduction basics pooling, including definition, principles, function Then, categorize summarize research preliminaries years. Next, explore potential applications, provides insightful insights promotion practice technology more fields. Furthermore, conduct comparative analysis evaluate performance on benchmark dataset, providing comprehensive understanding strengths weaknesses. Finally, systematically analyze opportunities prospective outlook directions.

Language: Английский

---