Projected $t$-SNE for batch correction

TL;DR

This paper introduces a new method for removing batch effects from t-SNE embeddings in high-dimensional biomedical data, improving interpretability while preserving biological signals.

Contribution

The authors propose a novel linear algebra and constrained optimization-based procedure to effectively correct batch effects in t-SNE visualizations, with fast computation.

Findings

Successfully removes multiple batch effects from t-SNE embeddings

Retains fundamental biological information such as cell types

Effective in real single-cell gene expression datasets

Abstract

Biomedical research often produces high-dimensional data confounded by batch effects such as systematic experimental variations, different protocols and subject identifiers. Without proper correction, low-dimensional representation of high-dimensional data might encode and reproduce the same systematic variations observed in the original data, and compromise the interpretation of the results. In this article, we propose a novel procedure to remove batch effects from low-dimensional embeddings obtained with t-SNE dimensionality reduction. The proposed methods are based on linear algebra and constrained optimization, leading to efficient algorithms and fast computation in many high-dimensional settings. Results on artificial single-cell transcription profiling data show that the proposed procedure successfully removes multiple batch effects from t-SNE embeddings, while retaining fundamental information on cell types. When applied to single-cell gene expression data to investigate mouse medulloblastoma, the proposed method successfully removes batches related with mice identifiers and the date of the experiment, while preserving clusters of oligodendrocytes, astrocytes, and endothelial cells and microglia, which are expected to lie in the stroma within or adjacent to the tumors.