Skip to content
BERMUDA (Batch Effect ReMoval Using Deep Autoencoders) is a novel transfer-learning-based method for batch correction in scRNA-seq data.
Python R
Branch: master
Clone or download
Latest commit 859516a Jul 17, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
R add data simulation Mar 20, 2019
pancreas add loss plot Mar 20, 2019
pbmc add loss plot Mar 20, 2019
BERMUDA.png update Apr 23, 2019
BERMUDA.py add BERMUDA code Mar 20, 2019
LICENSE Create LICENSE Jul 16, 2019
README.md
ae.py add BERMUDA code Mar 20, 2019
evaluate.py add BERMUDA code Mar 20, 2019
helper.py add BERMUDA code Mar 20, 2019
main_pancreas.py minor update Jul 17, 2019
main_pbmc.py minor update Jul 17, 2019
mmd.py add BERMUDA code Mar 20, 2019
pre_processing.py add BERMUDA code Mar 20, 2019

README.md

BERMUDA: Batch Effect ReMoval Using Deep Autoencoders

Tongxin Wang, Travis S Johnson, Wei Shao, Zixiao Lu, Bryan R Helm, Jie Zhang and Kun Huang

Codes and data for using BERMUDA, a novel transfer-learning-based method for batch-effect correction in single cell RNA sequencing (scRNA-seq) data.

BERMUDA

Dependencies

  • Python 3.6.5
  • scikit-learn 0.19.1
  • pyTorch 0.4.0
  • imbalanced-learn 0.3.3
  • rpy2 2.9.4
  • universal-divergence 0.2.0

Files

main_pancreas.py: An Example of combining two pancreas datasets
main_pbmc.py: An Example of combining PBMCs with pan T cells
R/pre_processing.R: Workflow of detecting clusters using Seurat and identifying similar clusters using MetaNeighbor
R/gaussian.R: Simulate data based on 2D Gaussian distributions
R/splatter.R: Simulate data using Splatter package

You can’t perform that action at this time.