CDState Resolves Malignant Cell Heterogeneity from Bulk Tumor RNA Sequencing Data.

CDState accurately deconvolves malignant cell states from bulk RNA-seq data, outperforming existing methods across five cancer types with 33 TCGA datasets.

• Why it matters: Understanding intratumor transcriptional heterogeneity is crucial for improving cancer treatment, but current single-cell methods are costly and limited in scale, creating a need for scalable alternatives.
• What they did: Researchers developed CDState, an unsupervised deconvolution method using nonnegative matrix factorization and cosine similarity, to infer malignant cell states and gene programs from bulk RNA-seq data.
• The result: CDState revealed recurrent gene programs linked to clinical features and identified genetic alterations as potential drivers of heterogeneity, enabling large-scale analysis of malignant cell states in cancer.