Validation_pseudobulk

Four pseudo-bulk RNA-seq datasets, simulated with SimBu [https://doi.org/10.1093/bioinformatics/btac499]. Basis for the simulation are two scRNA-seq datasets and true bulk RNA-seq datasets (uploaded and described in 10.6084/m9.figshare.25347757)Hao et al. [https://doi.org/10.1016/j.cell.2021.04.048] --&gt; finotello-simulation, hoek-simulationTabula Muris [https://doi.org/10.1038/s41586-018-0590-4] --&gt; chen-simulation, petitprez-simulationThe simulation setup with SimBu is:<br>For each true bulk sample n, we generated a simulation with 10 replicates. These simulated replicates all sampled 10,000 cells (with replacement) from the scRNA-seq datasets to have the same cell type composition as the bulk sample n (derived from FACS data); only the cells were used, that are present in both bulk and scRNA-seq dataset. Similarly, these pseudo-bulks also contain the same sequencing depth as bulk sample n. If we therefore have 8 bulk samples, we created 80 simulations (8 bulk samples * 10 replicates). Additionally, we scaled the expression of genes in each cell by scaling factors, derived by the total number of expressed genes per cell in the scRNA-seq reference dataset.<br><br>RNA-seq datasets are from different organisms and tissues:Chen: mouse, spleen [https://doi.org/10.1038/srep40508]Finotello: human, PBMC [https://doi.org/10.1186/s13073-019-0638-6]Hoek: human, PBMC [https://doi.org/10.1371/journal.pone.0118528]Petitprez: mouse, spleen [https://doi.org/10.1186/s13073-020-00783-w]