Tumour heterogeneity has a substantial impact on tumour progression and treatment response, yet bulk expression data obtained from clinical tumour samples obscure this complexity. Computational deconvolution methods can resolve cell-type-specific signals. This Review offers a practical guide for cancer researchers to select deconvolution methods and maximize the utility of bulk transcriptomic data.
- Yaoyi Dai
- Shuai Guo
- Wenyi Wang