There are five major sections in https://github.com/oxwang/fda_scRNA-seq
In preprocessing section, we provided script code to process bulk RNA-Seq and single cell RNA-Seq (scRNA-Seq) fastq data. For bulk RNA-Seq, one pipeline was provided. For scRNA-Seq, three pipelines were provided for 10x, C1_LLU, C1_FDA_HT, and iCELL8 data, respectively.
In pipeline_comp section, we provided R code to generate Figure 2 in our manuscript. The code is to compare the differences between any two pipelines for scRNA-Seq data.
In normalization section, we provided R code to generate sihouette values used for Figure 3 in our manuscript. The code is to generate sihouette values for the two gene count matrices with different sequencing depth 10k and 100k for seven datasets (10x_LLU, 10x_NCI, 10x_NCI_M, C1_LLU, C1_FDA_HT, iCELL8_PE, and iCELL8_SE) of HCC1395 and HCC1395BL samples. For each datasets in HCC1395 and HCC1395BL, there is an R code file.
In batch correction section, we provided R and python code to perform batch correction. The 'batch.R' represents the data loading code for four scenarios and Tian's data, and the code of seven batch correction. 'scanorama' folder represesnts the python code used for scanorama batch correction. 'bbknn.py' file represents the python code for bbknn batch correction. 'metrics_scores.R' represents the code to calculate alignment score, kBET score, and silhouette value.
In 'others' section, we provided the code to perform benchmarking assessment in section 5 & 6 in manuscript.