The goal of scrubletR is to run the scrublet in R.
You can install the released version of scrubletR from Github with:
remotes::install_github("Moonerss/scrubletR")-
R package: Matrix, reticulate, Seurat
-
Python module: scrublet
load("~/samples.RData")
res <- scrublet_R(seurat_obj = samples)
# Preprocessing...
# Simulating doublets...
# Embedding transcriptomes using PCA...
# Calculating doublet scores...
# Automatically set threshold at doublet score = 0.57
# Detected doublet rate = 0.1%
# Estimated detectable doublet fraction = 0.6%
# Overall doublet rate:
# Expected = 6.0%
# Estimated = 12.1%
# Elapsed time: 32.5 secondshead(res@meta.data)
# orig.ident nCount_RNA nFeature_RNA percent.mt
# AAACCCAAGAGTCACG-1 SeuratProject 5985 1844 7.552214
# AAACCCAAGCCGTCGT-1 SeuratProject 11923 2532 11.725237
# AAACCCAAGGACACTG-1 SeuratProject 1728 783 22.858796
# AAACCCACAAAGGTTA-1 SeuratProject 10676 2649 22.798801
# AAACCCACAGACGCTC-1 SeuratProject 2235 690 55.078300
# AAACCCACAGGAACCA-1 SeuratProject 14519 3402 4.979682
# doublet_scores predicted_doublets
# AAACCCAAGAGTCACG-1 0.01636424 FALSE
# AAACCCAAGCCGTCGT-1 0.03400332 FALSE
# AAACCCAAGGACACTG-1 0.22724604 FALSE
# AAACCCACAAAGGTTA-1 0.02451680 FALSE
# AAACCCACAGACGCTC-1 0.01955671 FALSE
# AAACCCACAGGAACCA-1 0.01955671 FALSE