Neurological Disorder Drug Discovery from Gene Expression with Tensor Decomposition

Background: Identifying effective candidate drug compounds in patients with neurological disorders based on gene expression data is of great importance to the neurology field. By identifying effective candidate drugs to a given neurological disorder, neurologists would (1) reduce the time searching for effective treatments; and (2) gain additional useful information that leads to a better treatment outcome. Although there are many strategies to screen drug candidate in pre-clinical stage, it is not easy to check if candidate drug compounds can also be effective to human. Objective: We tried to propose a strategy to screen genes whose expression is altered in model animal experiments to be compared with gene expressed differentially with drug treatment to human cell lines. Methods: Recently proposed tensor decomposition (TD) based unsupervised feature extraction (FE) is applied to single cell (sc) RNA-seq experiments of Alzheimer's disease model animal mouse brain. Results: Four hundreds and one genes are screened as those differentially expressed during A beta accumulation as age progresses. These genes are significantly overlapped with those expressed differentially with the known drug treatments for three independent data sets: LINCS, DrugMatrix, and GEO. Conclusion: Our strategy, application of TD based unsupervised FE, is useful one to screen drug candidate compounds using scRNA-seq data set.