miraliger ignoring reads with length < 18 nt

Hi Lorena,

At first, thanks for a wonderful tool!

I have recently encountered an interesting 'feature' of the `miraligner` tool - it seems it's ignoring reads with less than 18 nucleotides. When I check both mapped (first column of .mirna) and unamapped reads (.mirna.nomap) there are not reads with less than 18 nucleotides. Is there some reason why miraligner ignores such reads? miRBase for human (for example) contains several mature miRNAs with less than 18 nucleotides and I would like to include them as well. I am aware that those miRNAs have a good potential to be included in the database just because some misannotation but still I would like to include such reads in my analysis.

Thank,
Jan

Start by tracing how miraligner handles reads shorter than 18 nucleotides and how it writes the .mirna and .mirna.nomap outputs. Reproduce the behavior with short reads and verify that reads below 18 nt, including the relevant mature miRNAs, are represented in the appropriate output.

Written by the indexing model from the issue text.