scDR: Predicting Drug Response at Single-Cell Resolution

Genes (Basel). 2023 Jan 19;14(2):268. doi: 10.3390/genes14020268.


Heterogeneity exists inter- and intratumorally, which might lead to different drug responses. Therefore, it is extremely important to clarify the drug response at single-cell resolution. Here, we propose a precise single-cell drug response (scDR) prediction method for single-cell RNA sequencing (scRNA-seq) data. We calculated a drug-response score (DRS) for each cell by integrating drug-response genes (DRGs) and gene expression in scRNA-seq data. Then, scDR was validated through internal and external transcriptomics data from bulk RNA-seq and scRNA-seq of cell lines or patient tissues. In addition, scDR could be used to predict prognoses for BLCA, PAAD, and STAD tumor samples. Next, comparison with the existing method using 53,502 cells from 198 cancer cell lines showed the higher accuracy of scDR. Finally, we identified an intrinsic resistant cell subgroup in melanoma, and explored the possible mechanisms, such as cell cycle activation, by applying scDR to time series scRNA-seq data of dabrafenib treatment. Altogether, scDR was a credible method for drug response prediction at single-cell resolution, and helpful in drug resistant mechanism exploration.

Keywords: cancer; drug resistance; drug response; heterogeneity; scRNA-seq.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Gene Expression Profiling / methods
  • Humans
  • Melanoma*
  • RNA-Seq
  • Sequence Analysis, RNA / methods
  • Single-Cell Analysis* / methods

Grants and funding

This work was supported by the National Natural Science Foundation of China (62172213, 61872183) to W.J.