Prediction of associations between diseases and environmental factors of long and short non-coding RNAs

The lncRNA, miRNA, etc. molecules in the human body interact with each other, genes and environmental factors. The interactions can be described as a network structure. The interruption of the connections in the network structure for any reason or the formation of new connections in the network can cause various diseases. Testing in a laboratory whether there is a association between a molecule and a disease or a environmental factor is a time-consuming process that requires intensive human resources. It is thought that prediction of potential associations between non-linking molecules and environmental factors and diseases by methods to be developed will contribute to this process. Three methods have been proposed for this purpose. Two KATZ-based models have been developed in order to estimate the possible associations between LncRNA-environmental factors and miRNA-environmental factors. Another model has been developed based on stacked autoencoders and deep learning to estimate potential new associations between lncRNAs and diseases. The test results of the developed models show that they achieve more successful results in terms of AUC values against the models in the literature.