Authors

A McCluskey¹; T Kettlewell¹; TD Otto¹;  ¹ University of Glasgow , UK

Discussion

Bulk RNA Sequencing (RNA-seq) transformed our understanding for a plethora of organisms, being cheap and easy to set up. Single-cell RNA-seq (scRNA-seq), its successor, provides population resolution at a single-cell level, enabling study of organisms at unprecedented detail. One powerful case of this is trajectory analysis, where differentiating cells are ordered according to their transcriptional profile, and assigned a "pseudotime" value representing their developmental progress. Integrating RNA-seq with scRNA-seq leverages additional insights from the far cheaper and easier RNA-seq approaches, but it is challenging to perform this task in a reliable way. 

Here we present BLASE, which allows a researcher to map RNA-seq onto a precomputed scRNA-seq trajectory. BLASE is one of few tools that can integrate cell trajectories and RNA-seq time-series data, and outperforms other existing tools in our experiments with ground-truth bulk RNA-seq datasets. By identifying the position of an RNA-seq sample on pseudotime from scRNA-seq data, BLASE can provide a reciprocal advantage of enabling the use of ground-truth bulk RNA-sequences to assist with the annotation of scRNA-seq data. BLASE is available as a package on Bioconductor.

We also apply BLASE to the problem of understanding differences in developmental rate (i.e. the rate of transcriptional change, as opposed to the "growth rate" of a population of cells). By understanding where, for example, a wild-type (WT) and knock-out (KO) lie on a developmental trajectory, we can identify different developmental rates, and adjust downstream analyses (such as differential expression analysis (DEA), and gene ontology term (GO term) analysis) to minimise the downstream impact. We demonstrate this in the case of RNA-seq of a Plasmodium falciparum lrr5 KO line subjected to heat-shock in-vitro, showing increased synchronisation and increased developmental rate in the heat-shocked condition.