Interpretable visualization of single cell data using Janus autoencoders

Gokul Gowri, Philippa Anne Richter, Xiaokang Lun, Peng Yin

Keywords: single cell, multiomics, visualization, autoencoders, dimensionality reduction, interpretability

TL;DR: We develop a new autoencoder architecture that can find faithful low-dimensional representations of single cell data where the dimensions have interpretable biological meaning.

Abstract: The emergence of single-cell transcriptomics and proteomics approaches has resulted in a wealth of high-dimensional data that are challenging to interpret. Dimensionality reduction methods, such as UMAP and t-SNE, project data points onto a low-dimensional space that preserves cellular similarities from the high-dimensional measurement space. However, the projected dimensions typically have no interpretable biological meaning, and the relationships between measured biomolecular features are obscured completely. These limitations can be overcome by finding embeddings in which each dimension is a function of a distinct and biologically meaningful set of features. Here, we introduce Janus autoencoders, a novel neural network architecture capable of finding such low-dimensional embeddings by jointly optimizing multiple distinct one-dimensional embeddings of a dataset. We demonstrate the utility of Janus autoencoders for (1) visualizing multiomic data such that modality-specific contributions to cell state can be deconvolved and (2) visualizing mass cytometry data such that cell cycle effects can be distinguished from “true” cell state differences. Our initial demonstrations indicate that Janus autoencoders can visually represent relationships between cellular states and their underlying cellular features in multiple biological contexts.