pip install scalex

install the latest develop version

pip install git+https://github.com/jsxlei/scalex.git

or git clone and install

git clone git://github.com/jsxlei/scalex.git
cd scalex
python setup.py install

SCALEX is implemented in Pytorch framework.
SCALEX can be run on CPU devices, and running SCALEX on GPU devices if available is recommended.

SCALEX can both used under command line and API function in jupyter notebook

from scalex import SCALEX
adata = SCALEX(data_list, batch_categories)

Function of parameters are similar to command line options.
Output is a Anndata object for further analysis with scanpy.
data_list can be

• data_path, file format included txt, csv, h5ad, h5mu/rna, h5mu/atac, dir contains mtx
• list of data_paths
• Anndata
• list of AnnData
• above mixed

batch_categories is optional, name of each batch, will be range from 0 to N-1 if not provided

SCALEX.py --data_list data1 data2 dataN --batch_categories batch_name1 batch_name2 batch_nameN 

--data_list: data path of each batch of single-cell dataset, use -d for short

--batch_categories: name of each batch, batch_categories will range from 0 to N-1 if not specified

Output will be saved in the output folder including:

• checkpoint: saved model to reproduce results cooperated with option --checkpoint or -c
• adata.h5ad: preprocessed data and results including, latent, clustering and imputation
• umap.png: UMAP visualization of latent representations of cells
• log.txt: log file of training process

SCALEX.py --data_list <filename.h5ad>

SCALEX.py --data_list <filename.h5ad> --batch_name <specific_batch_name>

SCALEX.py --data_list <filename.h5ad> --profile ATAC

SCALEX.py --data_list <atac_filename.h5ad> --profile ATAC --impute True

SCALEX.py --data_list <filename.h5ad> --n_top_features features.txt

SCALEX.py --data_list <filename.h5ad> --processed

• --data_list
  A list of matrices file (each as a batch) or a single batch/batch-merged file.
• --batch_categories
  Categories for the batch annotation. By default, use increasing numbers if not given
• --batch_name
  Use this annotation in anndata.obs as batches for training model. Default: 'batch'.
• --profile
  Specify the single-cell profile, RNA or ATAC. Default: RNA.
• --min_features
  Filtered out cells that are detected in less than min_features. Default: 600 for RNA, 100 for ATAC.
• --min_cells
  Filtered out genes that are detected in less than min_cells. Default: 3.
• --n_top_features
  Number of highly-variable genes to keep. Default: 2000 for RNA, 30000 for ATAC.
• --outdir
  Output directory. Default: 'output/'.
• --projection
  Use for new dataset projection. Input the folder containing the pre-trained model. Default: None.
• --impute
  If True, calculate the imputed gene expression and store it at adata.layers['impute']. Default: False.
• --chunk_size
  Number of samples from the same batch to transform. Default: 20000.
• --ignore_umap
  If True, do not perform UMAP for visualization and leiden for clustering. Default: False.
• --join
  Use intersection ('inner') or union ('outer') of variables of different batches.
• --batch_key
  Add the batch annotation to obs using this key. By default, batch_key='batch'.
• --batch_size
  Number of samples per batch to load. Default: 64.
• --lr
  Learning rate. Default: 2e-4.
• --max_iteration
  Max iterations for training. Training one batch_size samples is one iteration. Default: 30000.
• --seed
  Random seed for torch and numpy. Default: 124.
• --gpu
  Index of GPU to use if GPU is available. Default: 0.
• --verbose
  Verbosity, True or False. Default: False.

Look for more usage of SCALEX

SCALEX.py --help