Skip to main content
Fig. 4 | Genome Biology

Fig. 4

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

Fig. 4

Effect of BUTTERFLY correction on gene expression. AC The log fold change in gene expression for the PBMC_V3_3 dataset between the downsampled data (at 1/10 number of reads) and the original data. A No correction. B Unseen molecules correction applied to every gene based on the data within the dataset. C The same correction as in B, but using additional data from six other 10x Chromium datasets for estimating the copies per UMI histograms per gene (Methods). D, E Two PBMC datasets generated with different technologies (EVALPBMC, 10x Chromium v2, and EVALPBMC_DS, Drop-Seq) were downsampled to varying depths and corrected for unseen molecules using BUTTERFLY. D The figure shows the change in correlation (CCC) between the two datasets at different prediction/downsampling points. Both datasets are downsampled/predicted to the same degree. E Change in gene expression from prediction/downsampling for three genes, for both datasets. The code to reproduce this figure is here: https://github.com/pachterlab/GRNP_2020/blob/master/notebooks/figure_generation/GenFig4AC_S23.ipynb (code for AC) and https://github.com/pachterlab/GRNP_2020/blob/master/notebooks/figure_generation/GenFig4DE.ipynb (code for D, E)

Back to article page