How to get good mapping for miRNA-seq data by removing rRNA/tRNA


I have 50 miRNA-seq samples from mammary tissue of cow. my data had no linker and adaptor. After quality trimming I aligned the. samples with bovine reference genome Bowtie with following command:
bowtie -m 10 -l 18 -n 1 -e 80 --best --strata -a
and I got mapping between 60% and 85% with average 78% but when I do mapping with mirbase this range decreases between 0.1 % to 65% with average 45%.
Data mapped with reference genome with ranged from 60 to 85 percent with an average of 78 percent, but when I map with bovine mirbase, this range decreases by 0.1 to 65 percent and by an average of 45 percent.

I wanted to check for tRNA,rRNA and snRNA contaminations. I did it by aligning samples to SILVA and rfam (fetched for tRNA,rRNA and snRNA) databases separately using Bowtie. But the mapping percent in average was less than 1 % and 3% respectively.

It may be necessary to point out that when I ran blast on a sample (with mapping percent of 0.01, 4, 0.7 and 65 percent with Silva, Rfam, bovine mirbase and reference genome respectively), about 50% mapped to bos taurus ribosomal RNA.
Could anybody give me some guidelines or provide some resources? I would appreciate if somebody can help me on it.



Source link