Fig. 1

Pan-cancer pre- and post-chemotherapy multi-omics analysis identified DUSP1 as a target enriched in minimal residual disease. A Schematic depicting the study design. We utilized 24 published datasets, consisting of 17 bulk RNAseq and array datasets, as well as 7 single cell RNA seq datasets. These datasets were obtained from 8 different cancer types and included a total of 1502 patients. BRCA: breast cancer, OV: ovarian cancer, NSCLC: non-small cell lung cancer, GBM: gliomablastoma, EAD: esophageal adenocarcinoma, ESCC: esophageal squamous cell carcinoma, COAD: colon adenocarcinoma, READ: rectal adenocarcinoma. B Bar graphs displaying the number of patients before and after chemotherapy in various bulk RNAseq and array datasets. C Bubble and bar graphs showing number of pre- and post-chemotherapy patients and cells across various single cell dataset. D Radar plot showing the percentage of genes that are significantly down-regulated (p < 0.05, log2(FC) < -1) and up-regulated (p < 0.05, log2(FC)- > 1) in patients who received chemotherapy compared to those who did not. The data is collected from 17 datasets representing 6 different types of cancer. BRCA: B1-B8, B1,GSE122630; B2, GSE123845; B3, GSE180280; B4, GSE191127; B5, GSE32072; B6, GSE32603; B7, GSE43816; B8, GSE87455; COAD: C1, GSE207194; EAD: E1, GSE165252; GBM: G1, GSE63035; OV: O1-O3, O1, GSE146965; O2, GSE16274; O3, GSE227100; READ: R1-R3, R1, GSE15781; R2, GSE233517; R3, GSE94104. E Genes found to be recurrently significantly differentially expressed between post- and pre-chemotherapy patients in multiple datasets. F Bubble plot showing the log2(FC) and -log10p of 35 most widely differentially expressed genes (DEGs) across 17 datasets. The marked down and up DEGs are related to cell cycle and ECM pathway respectively. G Violin plot comparing the expression changes of DUSP1 between patients samples before and after chemotherapy. Two sided Welch’s t-test was applied to calculate p values: (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). H Heat map shows the log2(FC) of immune infiltrates (calculated by MCPcounter method) between post- and pre-treatment patients across 17 datasets. Color represents log2(FC). I UMAP embedding overlaid with unsupervised cluster cell type annotations (left), treatment annotations (medium) and dataset annotations (right) of totally 323,664 cells integrating seven datasets. J The cell type composition in pre- and post-chemotherapy group of samples across various datasets and in the integrity. K Bubble plot showing the log2(FC) and -log10p of 35 most widely differentially expressed genes (DEGs) across 10 cell types and the integrity. L Comparison of DUSP1 expression between pre- and post-chemotherapy patient samples across fibroblast, T cell and mast cell. *p < 0.05