Hello!

I am new in NGS analysis, but in my new lab i had to replace my colleague in big NGS project related to miRNA-seq. He left me bunch of fastqs and mirbase 21 fasta file with prebuilt indexes ("old" attached file). Recently, i have downloaded mirbase 22 mature sequences as fasta file, to try to map my reads to a newer version of database. I have changed Us to Ts , and used bowtie-build to build new indexes)("new" attached file(renamed for upload). But when i run bowtie on the same samples with old file and the new one i had different results - ~35% unique mappers and ~4% multimappers with old mirbase 21 file ive got from colleague and exactly the opposite - 4% unique and ~35% multimappers - with the new file. As i said, i`m new in NGS analysis and i must be missing some simple processing step or something, that inverts my mapping results. Can you please explain what am i doing wrong and maybe give some links to good practical ngs course? Thank you!

https://quickfileshare.org/Z2y/[new]mirbase22.fa

https://quickfileshare.org/Z2z/[old]mirbase.fasta