Algorithm

Philosophy

Theoretical computer science

Curse of dimensionality

Operating system

Computer science

Simple (philosophy)

Graph

Snapshot (computer storage)

Inference

Epistemology

Artificial intelligence

Proceedings of the National Academy of Sciences

Significance Seeing a snapshot of individuals at different stages of a dynamic process can reveal what the process would look like for a single individual over time. Biologists apply this principle to infer temporal sequences of gene expression states in cells from measurements made at a single moment in time. However, the sparsity and high dimensionality of single-cell data have made inference difficult using formal approaches. Here, we apply recent innovations in spectral graph theory to devise a simple and asymptotically exact algorithm for inferring the unique dynamic solution under defined approximations and apply it to data from bone marrow stem cells.

Fundamental limits on dynamic inference from single-cell snapshots