I am doing an experiment of rnaseq. I have mapped Iontorrent reads (single end, 2 million reads, length average = 50) to a bacterial genome using bowtie2 and bbmap. Problem is that htseq does not attribute mapped reads to any gene, because all alignments have mapq scores =<1. Only if a utilize the option "-a 0" in htseq, I can see that the genes have mapped reads. I was reading throughout biostars that a mapq threshold should be set to 5, at least. Visually, I perceived that most of the alignments were on rRNA gene, demonstrating that the ribosomal reduction was not efficient. But, anyway, most of the genes contained aligned reads. So, is it acceptable to set a mapq score to 0? If not, what are my possibilities? How could I improve mapq scores?