A representative analysis can be performed using the RNA-seq data SRA accession SRR319229 that was the Wold dataset that was used in the original analysis of bias.
This can be done using a prealigned bam file that can be downloaded from here. Alternatively the bam file can be created by downloading and aligning the raw read data as follows:
The raw data can be downloaded from here. Its recommended that this is done by installing the SRA toolkit from here using the command:
This can then be aligned to a reference genome using hisat2 which can be downloaded from here and then aligned using the following command:
hisat2 –p 10 -x grcm38_tran/genome_tran -U SRR3192229.fastq -S SRR3192229.sam
samtools view -b SRR3192229.sam > SRR3192229.bam
where genome_tran is the root name of the reference genome that can be downloaded from here
After the bam file has either been downloaded or created, normalisation of these data can then be performed by LiBiNorm using the following command:
LiBiNorm count -p 3 -f -s no -u wold -c counts.txt SRR3192229.bam Mus_musculus.GRCm38.80.gtf
This specifies that all the models are evaluated but that model BD is nevertheless used to correct for the bias. Auxilliary output files all have the prefix wold and the expression data is output to counts.txt. Three threads are used.
This requires a reference annotation that can be downloaded and uncompressed from here. This will produce the output data files that are described here
The graphical represntations as shown here can be produced using the command:
Rscript LiBiNormPlot.R wold