Code and scripts for reproducing figures of the BRB-seq paper
RNA-seq raw data (fastq files) are accessible in ArrayExpress under accession numbers E-MTAB-6469, E-MTAB-6984 and E-MTAB-7524. TruSeq datasets on LCL GBR samples from the 1000 Genomes project were downloaded from ArrayExpress E-MTAB-3656 and E-GEUV-1 respectively. The public SCRB-seq datasets were from Hafner et al. (GSE80297); Cacchiarelli et al. (GSE62777); Kilens et al. (PRJEB18663) and Xiong et al. (GSE98432).
The sequencing reads were aligned to the Ensembl r87 gene annotation of the hg38 genome using STAR (version 2.5.3a). Then, using the BRB-seqTools suite, we performed simultaneously the second demultiplexing and the count of reads/transcripts (UMI) per gene from the R1 FASTQ and the aligned R2 BAM files. This generated two count matrices (reads and UMI) that were used for further analyses. In the case of TruSeq, count matrices were generated with HTSeq (version 0.9.1).
Note: Count matrices used in the scripts were deposited on this GitHub page, in the data folder
The BRB-seqTools suite is implemented in Java. The version used in the manuscript (source code and tool) is permanently available under https://doi.org/10.5281/zenodo.2552405. It supports all the required post-sequencing tasks (alignment, counts) up until the generation of the read/UMI count matrix.
Of note: Now, it's rather recommended to use STARsolo for performing the same tasks