Prediction of Drug-Target Interactions with High-Quality Negative Samples and A Network-Based Deep Learning Framework.

Identification of drug-target interactions (DTIs) plays a crucial role in drug discovery. Compared to traditional experimental methods, computer-based methods for predicting DTIs can significantly reduce the time and financial burdens of drug development. In recent years, numerous machine learning-based methods have been proposed for predicting potential DTIs. However, a common limitation among these methods is the absence of high-quality negative samples. Moreover, the effective extraction of multisource information of drugs and proteins for DTI prediction remains a significant challenge. In this paper, we investigated two aspects: the selection of high-quality negative samples and the construction of a high-performance DTI prediction framework. Specifically, we found two types of hidden biases when randomly selecting negative samples from unlabeled drug-protein pairs and proposed a negative sample selection approach based on complex network theory. Furthermore, we proposed a novel DTI prediction method named HNetPa-DTI, which integrates topological information from the drug-protein-disease heterogeneous network and gene ontology (GO) and pathway annotation information of proteins. Specifically, we extracted topological information of the drug-protein-disease heterogeneous network using heterogeneous graph neural networks, and obtained GO and pathway annotation information of proteins from the GO term semantic similarity networks, GO term-protein bipartite networks, and pathway-protein bipartite network using graph neural networks. Experimental results show that HNetPa-DTI outperforms the baseline methods on four types of prediction tasks, demonstrating the superiority of our method. Our code and datasets are available at https://github.com/study-czx/HNetPa-DTI.

GCNA-Cluster: A Gene Co-Expression Network Alignment to Cluster Cancer Patients Algorithm for Identifying Subtypes of Pancreatic Ductal Adenocarcinoma.

Cancer heterogeneity makes it necessary to use different treatment strategies for patients with the same pathological features. Accurate identification of cancer subtypes is a crucial step in this approach. The current studies of pancreatic ductal adenocarcinoma (PDAC) subtypes mainly focus on single genes and ignore the synergistic effects of genes. Here we proposed a network alignment algorithm GCNA-cluster to cluster patients based on gene co-expression networks. We constructed weighted gene co-expression networks for patients and aligned the networks of two patients to estimate the similarity of patients and their cancer subtypes. A scoring function is defined to measure the network alignment result and the score can indicate the similarity between patients. Then, the patients are clustered based on their similarities. We validated the accuracy of the algorithm on the GEO-PDAC dataset with real labels, and the experimental results show that the GCNA-cluster algorithm has better results than classical cancer subtyping algorithms. In addition, the GCNA-cluster algorithm applied to the TCGA-PDAC dataset identified two subtypes based on the Silhouette Coefficient. Biomarkers identified for the PDAC subtypes hint to cell growth, cell cycle or apoptosis as targets for new therapeutic strategies.

Identification of subtype specific biomarkers of clear cell renal cell carcinoma using random forest and greedy algorithm.

Suitable biomarkers can be good indicator for cancer subtype. To find biomarkers that can accurately distinguish clear cell renal cell carcinoma (ccRCC) subtypes, we first determined ccRCC subtypes based on the expression of mRNA, miRNA and lncRNA, named clear cell type 1 (ccluster1) and 2 (ccluster2), using three unsupervised clustering algorithms. Besides being associated with the expression pattern derived from the single type of RNA, the differences between subtypes are relevant to the interactions between RNAs. Then, based on ceRNA network, the optimal combination features are selected using random forest and greedy algorithm. Further, in survival-related sub-ceRNA, competing gene pairs centering on miR-106a, miR-192, miR-193b, miR-454, miR-32, miR-98, miR-143, miR-145, miR-204, miR-424 and miR-1271 can also well identify ccluster1 and ccluster2 with prediction accuracy over 92%. These subtype-specific features potentially enhance the accuracy with which machine learning methods predict specific ccRCC subtypes. Simultaneously, the changes of miR-106 and OIP5-AS1 affect cell proliferation and the prognosis of ccluster1. The changes of miR-145 and FAM13A-AS1 in ccluster2 have an effect on cell invasion, apoptosis, migration and metabolism function. Here miR-192 displays a unique characteristic in both subtypes. Two subtypes also display notable differences in diverse pathways. Tumors belonging to ccluster1 are characterized by Fc gamma R-mediated phagocytosis pathway that affects tissue remodeling and repair, whereas those belonging to ccluster2 are characterized by EGFR tyrosine kinase inhibitor resistance pathway that participates in regulation of cell homeostasis. In conclusion, identifying these gene pairs can shed light on therapeutic mechanisms of ccRCC subtypes.

Temporal variation and spatial scale dependency of the trade-offs and synergies among multiple ecosystem services in the Taihu Lake Basin of China.

Grasping the temporal-spatial characteristics of interactions and spatial scales of multiple ecosystem services is the foundation for sustainable ecosystem management. Eight ecosystem services, including crop production, freshwater supply, aquatic production, net primary production, soil conservation, water retention, flood regulation, and forest recreation were measured at the 1-km2 pixel scale in the Taihu Lake Basin (TLB) of China from 1990 to 2010. Furthermore, we quantified the trade-offs and synergies of services at different periods of urbanization and across the 1-km2 pixel scale and the county scale. We aim to find which ecosystem services interactions temporally vary and depend on spatial scale. Our results found that: 1). Tremendous amount of cultivated lands were converted to construction land, and rapidly shrank from 1990 to 2010. 2). Determined by land use, different ecosystem services had spatial heterogeneity of their strength. Ecosystem services hot spots experienced an increasing trend while cold spots showed a trend of decreasing first and then increasing from 1990 to 2010. 3). Trade-offs between provisioning services and regulating services at the 1-km2 pixel scale changed over time. There was a new synergy between freshwater supply and aquatic production at the 1-km2 pixel scale in 2010 with the human demand. 4). From 1990 to 2010, the changes of provisioning services led to trade-offs among provisioning services, regulating services and cultural services at two scales. Taking temporal variation and scale dependence into account, this research is helpful to the delineation of "Ecological Conservation Redline" and implement the project of "Grain for Green". We also provide suggestions for maintaining ecosystem services with economic growth in China's Yangtze River Economic Belt for land use policies and decision making.