I have human miRNA-seq data. In the past I have aligned these data using Bowtie2 to the reference genome (Homo_sapiens.GRCh38.dna.toplevel.fa), and I have subsequently performed counting using featureCounts with the annotation file hsa.gff3 from miRBase. . Now I have aligned the reads to the mature miRNA from miRBase (mature.fa), but when I look at the resulting bam files, the reads have a flag of 4 (segment unmapped) and there is no position information in mature.fa.
So can I use mature.fa as the reference for alignment and if so how is counting performed?