Code to reproduce figures and tables in paper https://www.nature.com/articles/s41467-021-25393-x
This code was written in the NIH Integrated Data Analysis Portal (NIDAP), a user interface developed on the Foundry Platform (Palantir Technologies).
Code Authors: Margaret Cam, Thomas Joshua Meyer, Christian Sidak, Matthew Angel, Richard Finney, Jing Bian.
Quality_Control.R > Color_by_Genes.R > ModScore.R > DEG_l2p.R
Note: unused R objects are deleted after each section/step to reduce memory load.
Step 1: Filter & Generate Quality Control Plots from h5 files
Step 2: Generate Post-filter Data
Step 3: PCA & Normalization
Step 4: Combine and Renormalize
Produces figures 3b, 5a, 5e, 5g, 7e, 7m, S5b, S5c
Step 5: ModScore and Cell Classification (produces figure 5b) Note: This code replaces gene expression information of irrelevant genes (e.g. Vamp4 and Vash2) with information from negative transcriptional markers (e.g. Cd4_neg, Sell_low). See Supplementary Table 3: Marker genes used for cell type identity by scRNAseq.
Step 6: Dotplot and Contingency Table (produces figures 3a, S3d, table S4)
Step 7: DEG Analysis
Step 8: List to Pathway Visualisation (produces figures 4a, 5f, 6b, 6d, 7d, S7k, and tables S5, S6, S8, S9, S10, S11)
Note: List to Pathway Visualisation (l2p) requires l2p_0.0-1 (see attached tar.gz file). Github link: https://github.com/CCBR/l2p