Skip to main content
. 2021 Oct 8;81(23):5813–5817. doi: 10.1158/0008-5472.CAN-21-0899

Figure 1.

Figure 1. Overview of workflow for mutational signature discovery/prediction, visualization, and analysis. A, Workflow allows for loading and combining data from multiple sources, de novo discovery of signature and exposures, and prediction of exposures from existing signatures. B, The same sample exposures are plotted (subset to top samples), proportional exposures (signature exposures sum to 1), split up by tumor type, and split up by signature. C, UMAP. Tumors in a UMAP can be colored by annotations such as tumor type, clustering by methods such as K-means, or levels of exposure for each signature. D, Downstream analysis tools include automated comparison to COSMIC signatures, heatmaps, which can be used to show the relative levels of signature exposures in samples along with sample annotations, and differential analysis of exposures between groups of tumors.

Overview of workflow for mutational signature discovery/prediction, visualization, and analysis. A, Workflow allows for loading and combining data from multiple sources, de novo discovery of signature and exposures, and prediction of exposures from existing signatures. B, The same sample exposures are plotted (subset to top samples), proportional exposures (signature exposures sum to 1), split up by tumor type, and split up by signature. C, UMAP. Tumors in a UMAP can be colored by annotations such as tumor type, clustering by methods such as K-means, or levels of exposure for each signature. D, Downstream analysis tools include automated comparison to COSMIC signatures, heatmaps, which can be used to show the relative levels of signature exposures in samples along with sample annotations, and differential analysis of exposures between groups of tumors.