Just as detectives use clues to solve mysteries, researchers use computational methods to uncover the genes behind diseases like Alzheimer’s and Parkinson’s. In this study, scientists introduced a prediction method called preserving structure network embedding (PSNE). They constructed a complex network by combining disease-gene associations, human protein networks, and disease-disease associations. By extracting low-dimensional features from the network, they were able to uncover new disease-gene relationships. Compared to other methods, PSNE proved to be more effective in predicting disease genes. The researchers then applied PSNE to predict potential pathogenic genes for age-associated diseases. To confirm their predictions, they compared them with existing literature. As a result, they identified several high-confidence potential pathogenic genes for Alzheimer’s and Parkinson’s. This research provides an exciting glimpse into the intricate world of disease gene prediction and may pave the way for future experimental discoveries. To dive deeper into their findings, check out the original paper!

Many diseases, such as Alzheimer’s disease (AD) and Parkinson’s disease (PD), are caused by abnormalities or mutations of related genes. Many computational methods based on the network relationship between diseases and genes have been proposed to predict potential pathogenic genes. However, how to effectively mine the disease-gene relationship network to predict disease genes better is still an open problem. In this paper, a disease-gene-prediction method based on preserving structure network embedding (PSNE) is introduced. In order to predict pathogenic genes more effectively, a heterogeneous network with multiple types of bio-entities was constructed by integrating disease-gene associations, human protein network, and disease-disease associations. Furthermore, the low-dimension features of nodes extracted from the network were used to reconstruct a new disease-gene heterogeneous network. Compared with other advanced methods, the performance of PSNE has been confirmed more effective in disease-gene prediction. Finally, we applied the PSNE method to predict potential pathogenic genes for age-associated diseases such as AD and PD. We verified the effectiveness of these predicted potential genes by literature verification. Overall, this work provides an effective method for disease-gene prediction, and a series of high-confidence potential pathogenic genes of AD and PD which may be helpful for the experimental discovery of disease genes.

Read Full Article (External Site)