• Demonstration code of the old Tensorflow code (the demo_code in this repository loads the CORTEX dataset)
• Reproducing the experiments of the paper "Deep generative modeling for single-cell transcriptomics" https://www.nature.com/articles/s41592-018-0229-2
• Current version in PyTorch is available at https://github.com/YosefLab/scVI

romain [underscore] lopez [at] berkeley [dot] edu

• CORTEX

Original file expression_mRNA_17-Aug-2014.txt can be dowloaded here https://storage.googleapis.com/linnarsson-lab-www-blobs/blobs/cortex/expression_mRNA_17-Aug-2014.txt

• PBMC

An additional R script to extract the quantity of interest from SCONE objects is needed (additional/PBMC.R) scVI_scone.rda, scone_all_wposcon_extendim_biostrat2_q.rda, molecule_qc_8k.txt, molecule_qc_4k.txt and gene_info.csv (containing the merged p-values from the microarray experiments) are in additional/data.zip

• BRAIN LARGE

An additional Jupyter Notebook to format the data for scVI is needed (additional/BRAIN-LARGE.ipynb) 1M_neurons_filtered_gene_bc_matrices_h5.h5 can be downloaded here http://cf.10xgenomics.com/samples/cell-exp/1.3.0/1M_neurons/1M_neurons_filtered_gene_bc_matrices_h5.h5

• RETINA

An additional R script, based on the original analysis is needed (additional/RETINA.R) We modified the file class.R so that the gene filtering also apply to the raw matrix, use ours in this repo (additional/class.R) The file bipolar_data_Cell2016.Rdata can be downloaded here https://github.com/broadinstitute/BipolarCell2016 The file ClustAssignFile.txt was obtained from the author, uploaded in additional/data.zip

• HEMATO

Original files bBM.raw_umifm_counts.csv can be downloaded here https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM2388072 bBM.spring_and_pba.csv and bBM.filtered_gene_list.paper.txt were sent by authors, uploaded in additional/data.zip

• CBMC

Original files GSE100866_CBMC_8K_13AB_10X-RNA_umi.csv, GSE100866_CBMC_8K_13AB_10X-ADT_umi.csv, GSE100866_CBMC_8K_13AB_10X-ADT_clr-transformed.csv can be downloaded here https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE100866

• BRAIN SMALL

Original files genes.tsv, barcodes.tsv, matrix.mtx and clusters.csv can be downloaded here https://support.10xgenomics.com/single-cell-gene-expression/datasets/2.1.0/neuron_9k molecule_qc.txt was extracted from the original file http://cf.10xgenomics.com/samples/cell-exp/2.1.0/neuron_9k/neuron_9k_molecule_info.h5 using the additional python script additional/molecule_info.py (credit David Detomaso)

• Python 2.7.12 (version tested) or compatible
• Tensorflow 1.3.0 (version tested) or compatible

• Clone the github repository, install the dependencies and call functions from the modules scVI
• Install time (< 10 min)

• simply follow the given notebook
• output should match the paper
• run time should be less than 10 min on any hardware (including CPU, CORTEX is a small dataset)

• Follow the analysis notebooks