Bowtie to analyze miRNA

[gkuffel]

Hi everyone,

I have 50 bp single end miRNA reads from mouse generated on the MiSeq. So far I have trimmed adapters using Cutadapt and now I am mapping to the whole genome using Bowtie. After mapping I plan to count features in the BAM file using HTSeq to count the miRNAs. I am having a hard time understanding why I should allow multiple mappings for a given read. The command I am using is below:

bowtie --wrapper basic-0 --wrapper basic-0 -n 0 -l 15 -e 99999 --sam --best

My issue is this, eventhough I am getting great alignments (around 90% of my reads map), when I count the features only about 10% of the reads are miRNAs?

When I change my bowtie parameter to include -k 200 which allows each read to map up to 200 different places in the genome the number of miRNAs obviously increases. If you are an expert in miRNA analysis your expertise would be greatly appreciated here.

[sbarberan]

To reduce problems with multiple mappings for miRNAs my approach is always align against miRBase first, and then the unaligned reads against the genome.

Why do you use bowtie and not bowtie2?

This paper convinced me of using bowtie2:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4931105/