Hello. I'm using Rsubread to align short stranded Illumina RNA seq reads to the yeast (S. cerevisiae) genome). The BAM file generated using subjunc shows quite a large number of reads that are erroneously mapped across very long distances, considered introns. One of the mapped ends often corresponds to stretches of "A". I was wondering if there is an option that could limit the size of the detected introns or filter out reads that were partially mapped in low complexity genomic regions ?

Thank you very much.

The test command used (using indexed yeast genome file from Ensembl and the corresponding annotation file):

subjunc(index="yeast110", readfile1="RAW/my.fastq.gz", 
                     output_file = "BAM/my.bam", 
                     output_format = "BAM",
                     nthreads = 8,
                     sortReadsByCoordinates = TRUE,
                     annot.ext = "Saccharomyces_cerevisiae.R64-1-1.110.gtf.gz",
                     isGTF = T,
                     useAnnotation = T)

EDIT: since STAR has a specific parameter that allows to define maximum acceptable intron size, I'm switching back to it. I liked the idea of using only R for the whole workflow, but it's not a big problem to switch to a few bash commands.