Beyondcell is a computational methodology for identifying tumour cell subpopulations with distinct drug responses in single-cell RNA-seq and Spatial Transcriptomics data.
Beyondcell provides an innovative method to discover drugs that operate differently among single cell clusters. As indicated, switch points closer to 0 represent susceptibility to the drug, while values closer to 1 suggest resistance to the drug. However, the results are ranked from high to low switch point. As seen in the function frank() (Rank.R script, line 215), drugs are ordered by descending values of switch point followed by residuals mean, which (i) leaves drugs with predicted higher resistance on top and (ii) drugs with negative residuals and thus low group enrichment are ranked higher that drugs from the next switch point tier. I'd suggest reversing the sign of ranking by switch point or updating as considered appropriate.
Beyondcell provides an innovative method to discover drugs that operate differently among single cell clusters. As indicated, switch points closer to 0 represent susceptibility to the drug, while values closer to 1 suggest resistance to the drug. However, the results are ranked from high to low switch point. As seen in the function
frank()(Rank.R script, line 215), drugs are ordered by descending values of switch point followed by residuals mean, which (i) leaves drugs with predicted higher resistance on top and (ii) drugs with negative residuals and thus low group enrichment are ranked higher that drugs from the next switch point tier. I'd suggest reversing the sign of ranking by switch point or updating as considered appropriate.Best regards.