Rsubread for miRNA

[inah]

For micro RNA (miRNA) data, the following aligners are recommended specifically for these short sequences:
MicroRazerS (www.seqan.de/projects/microrazers/)
mrFAST (mrfast.sourceforge.net/)
mrsFAST (mrsfast.sourceforge.net/Home)
PatMaN (bioinf.eva.mpg.de/patman/)
Does anyone know how the Rsubread align function compares to these? Has anyone performed any comparisons? I use Rsubread for RNAseq and it would be convenient to use it also for miRNAseq, but I am a little concerned and wonder whether I need to invest time in conducting some comparisons.

I have noticed one potential problem with Rsubread featureCounts function when applied to miRNAseq: When I use the annotation file from mirBase (hsa.gff3) instead of the built-in annotation or the ensembl GTF file, then the Gene IDs in the counts (rownames) and annotation output from Rsubread-featureCounts are all NA (see code below).

counts_TH14_uniqtrue_annotMirBmature.out <- featureCounts(files=mapped.flist,
annot.inbuilt="hg38", chrAliases=NULL,
# use mirBase GTF file and feature = miRNA (mature miRNA)
annot.ext="/home/inah/Rsubread_miRNA/RefGTF/hsa.gff3",
isGTFAnnotationFile=TRUE,
GTF.featureType="miRNA", GTF.attrType="miRNA", useMetaFeatures=FALSE, ...

Many thanks, Ina

[Elfangor]

For miRNA I used bowtie1 with --best --strata and bowtie2 with --very-sensitive-local options. I found around a 90% alignment score between my reads and reference genome.

Based on Ziemann et al 2016. "Evaluation of miRNA alignment techniques".

This is the code I used for featureCounts:

featureCounts -t miRNA -g ID -f -s 1 -O -T 16 -a /smallRNA.gtf -o /output_dir n.bam 2>featureCounts.log