Fig. 5

Inferring causality using pseudotime analysis and RNA velocity. a In this hypothetical example, when determining the relationship between gene X and gene Y, no correlation would be observed (top). However, the relationship between both genes may be masked by a time delay and correcting for this time delay might reveal a clear correlation between the expression of gene X and gene Y (bottom). b To identify the length of a time delay, the cells can be ordered along pseudotime, that is, an inferred timeline based on the variable gene expression states of single cells measured at a single moment in time (top). RNA velocity [86], a readout that exploits the unidirectional character of splicing, allows the prediction of the future state of a cell and its genes (bottom). As such, the correct ordering of cells can be validated using RNA velocity. Plotting gene expression against pseudotime shows that the expression of gene X is following the expression of gene Y. From this, it can be deduced that gene Y is regulating gene X, and not the other way around