Embedding to Reference t-SNE Space Addresses Batch Effects in Single-Cell ClassificationPoličar, Pavlin Gregor (Avtor) Zupan, Blaž (Mentor) batch effectsembeddingt-SNEvisualizationsingle-cell transcriptomicsdata integrationdomain adaptationDimensionality reduction techniques, such as t-SNE, can construct informative visualizations of high-dimensional data. When working with multiple data sets, a straightforward application of these methods often fails; instead of revealing underlying classes, the resulting visualizations expose data set-specific clusters. To circumvent these batch effects, we propose a principled embedding procedure that enables the addition of new data points into existing t-SNE embeddings. We provide an open-source implementation of the proposed method and demonstrate the utility of our approach with an analysis of six recently published single-cell gene expression data sets containing up to tens of thousands of cells and thousands of genes. We present surprising evidence that our computationally more direct procedure solves the batch effect problem, one of the core challenges in the analysis of gene expression data, and enables the reuse of t-SNE embeddings, paving the way for interpretable visualizations of high-dimensional data sets.20192019-08-30 13:20:04Magistrsko delo/naloga109337sl