Thread by @thatdnaguy

#ASHG21 FW: Frank Wessely.
Understanding cell-type-specific drug effects while controlling for genotype.

#ASHG21 FW: there are databases of gene expression responses to perturbation. Such as Cmap and CLUE.

#ASHG21 FW: drug perturbation in cellular context. 1,680 bulk RNA-Seq samples from 7 hour single dose of 1/12 compounds. 14 cell types. Used 3' Quant-Seq.

#ASHG21 FW: Cell type main driver of variance across the study. Further some separation by layer and level of differentiation.

#ASHG21 FW: Cell line identity, i.e. genotype, drives transcriptional variability. [hmm. Are these same cell types derived from different individuals? I'm not sure]

#ASHG21 FW: Half of tested compounds induce weak response of <50 DE genes. Observed variability similar to what is observed in connectivity map (CMap2) for the same compound.

#ASHG21 FW: Highly variable response to the same compound amongst different cells.

#ASHG21 FW: Response to treatment varies between cells. MCF cancer cell line showed the least responsiveness. Progenitor cells showed response to all compounds (though some are small). Undifferentiated cells share common response with differentiated counterparts.

#ASHG21 FW: majority of compounds share a small core set of perturbed genes between multiple cell types.

#ASHG21 FW: The core set can be as small as a single gene, which may be the target gene for the compound. Directionality of shared gene changes is usually the same in all cells that share it.

#ASHG21 FW: Conclusion is that the responses are variable generally among cell types, but often have a core change of a few genes that are shared with all. Progenitors often show similar responses to the differentiated cells derived from them.