I'm aligning sequencing data to the human mitochondrial genome with Bowtie2. To handle the fact that the genome is circular, I'm doing two alignments (in which one has a shifted origin) as aligners treat the genome as linear. My goal was to replace the reads from one to the other where the origin is. However, when I do this, regions that should have the same number of reads in each alignment are significantly different. Further, the total number of reads that align is also VERY different. Does anyone have any suggestions on what might be going wrong or what to do?