Skip to main content
Fig. 5 | Genome Biology

Fig. 5

From: BUTTERFLY: addressing the pooled amplification paradox with unique molecular identifiers in single-cell RNA-seq

Fig. 5

Applications of amplification bias correction. A, B. Batch correction of datasets with different sampling depths. The figures show a mix of the T cell-like cells from the PBMC_NG and PBMC_NG_2 datasets, as determined by clustering, in UMAP space. A Uncorrected data. B Data corrected using binomial downsampling of PBMC_NG_2. CH Differences in amplification across clusters. C Joint CU histogram of the T cell clusters T1-T4 for the ALDH2 gene. D Joint CU histogram of the monocyte clusters M1-M2 for the ALDH2 gene. E Gene expression for the ALDH2 gene plotted against copies per UMI across clusters. F Corrected gene expression (using Preseq DS, predicted to 1020) for the ALDH2 gene plotted against copies per UMI across clusters. G The gene expression of the 50 genes with most variation across clusters plotted against copies per UMI, where each point represents a unique cluster and gene. H Similar to G, but for corrected data (using Preseq DS, predicted to 1020). No notebooks are available for the code to produce these figures, but the code is available as stated in the “Availability of data and materials” section

Back to article page