I am comparing different aligners for miRNA data analysis. I used miRNA precursor sequences (70-100 nt in length) to create an index file called "hairpin.gaa". Then, I mapped all clean miRNA-seq data (about half million reads, length is about 21 nt) to these precursor sequences, using the following command:
bowtie hairpin.gaa -q seq.fastq -v 1 -a --best --strata --norc -S out.sam
The result is very surprising. I only get 200 alignments in total. If I change "-v 1" into "-v 3", I can get about 3,000 alignments in total. In contrast, when we used GSNAP tool, we are able to get decent alignment for ~20% of total reads.
Here, I just want to make sure that the parameters I used are appropriate for miRNA analysis. Also, I want to know whether I need to use special parameters for bowtie-build to create the index file for this case or not?
Any comment or suggestion will be highly appreciated.