Get read counts for reads overlapping exon-intron junction from STAR-output

Question

[Palgrave]

Hi, I am looking for a method to get counts per gene for the spliced read output from the *.tab files from STAR. Is there a way to do this directly in STAR?

Answer 1

[caggtaagtat]

Could you maybe use the STAR option --quantMode GeneCounts? This gives you a count matrix with read counts per gene

Edit:

sorry I read you question wrong. You can use the generated bam file from STAR und convert it back to a FASTQ e.g. with bedtools and use it for pseudomapping with salmon or kalisto.

[Palgrave]

Which bam file should I use? I get to mapping files: *Aligned.out.sam and *Chimeric.out.sam

[caggtaagtat]

You would use the Aligned.out.sam file. But you have to select uniquely mapped reads first from it. That could look something like this with samtools:

/samtools/bin/samtools view -q 255 -b Unfiltered_file.bam > filtered_file.bam

Otherwise you get reads, which mapped to multiple loci.

[Palgrave]

But how do I identify only the reads overlapping a splice junction from the sam file?

[caggtaagtat]

STAR outputs only gapped reads