Abstract:MOTIVATION:Single-cell technologies have recently gained popularity in cellular differentiation studies regarding their ability to resolve potential heterogeneities in cell populations. Analyzing such high-dimensional single-cell data has its own statistical and computational challenges. Popular multivariate approaches are based on data normalization, followed by dimension reduction and clustering to identify subgroups. However, in the case of cellular differentiation, we would not expect clear clusters to be present but instead expect the cells to follow continuous branching lineages.RESULTS:Here, we propose the use of diffusion maps to deal with the problem of defining differentiation trajectories. We adapt this method to single-cell data by adequate choice of kernel width and inclusion of uncertainties or missing measurement values, which enables the establishment of a pseudotemporal ordering of single cells in a high-dimensional gene expression space. We expect this output to reflect cell differentiation trajectories, where the data originates from intrinsic diffusion-like dynamics. Starting from a pluripotent stage, cells move smoothly within the transcriptional landscape towards more differentiated states with some stochasticity along their path. We demonstrate the robustness of our method with respect to extrinsic noise (e.g. measurement noise) and sampling density heterogeneities on simulated toy data as well as two single-cell quantitative polymerase chain reaction datasets (i.e. mouse haematopoietic stem cells and mouse embryonic stem cells) and an RNA-Seq data of human pre-implantation embryos. We show that diffusion maps perform considerably better than Principal Component Analysis and are advantageous over other techniques for non-linear dimension reduction such as t-distributed Stochastic Neighbour Embedding for preserving the global structures and pseudotemporal ordering of cells.AVAILABILITY AND IMPLEMENTATION:The Matlab implementation of diffusion maps for single-cell data is available at https://www.helmholtz-muenchen.de/icb/single-cell-diffusion-map.CONTACT:fbuettner.phys@gmail.com, fabian.theis@helmholtz-muenchen.deSUPPLEMENTARY INFORMATION:Supplementary data are available at Bioinformatics online.

Functional Diffusion Maps

Recent Developments in Complex and Spatially Correlated Functional Data

Diffusion Maps for High-Dimensional Single-Cell Analysis of Differentiation Data

Functional optimal transport: map estimation and domain adaptation for functional data

Manifold Diffusion Fields

Continuous-Time Functional Diffusion Processes

Multivariate Functional Principal Component Analysis for Data Observed on Different (Dimensional) Domains

Functional Data Analysis: An Introduction and Recent Developments

Model-Based Clustering and Classification of Functional Data

A Functional approach for Two Way Dimension Reduction in Time Series

Optimal Linear Discriminant Analysis for High-Dimensional Functional Data

G-invariant diffusion maps

Diffusion Component Analysis: Unraveling Functional Topology in Biological Networks

Review of Functional Data Analysis

Functional Nonlinear Learning

Diffusion map for clustering fMRI spatial maps extracted by independent component analysis

Functional data analysis using deep neural networks

Modeling Diffusion in Functional Materials: From Density Functional Theory to Artificial Intelligence

Bayesian Analysis of Multidimensional Functional Data

Inverse regression for spatially distributed functional data