miRNA counting - what am I missing?

[pmiguel]

That's not quite what I am wondering about.

I have a concrete miRNA, say, hsa-miR-92a. I see certain number of reads mapping there, about 16,000, of which about 6,000 are unique mappers. Each of those reads has a unique Illumina name, and starts on the right spot (beginning of mature 3p-miRNA).

At the same time, IGV and featureCounts recognize only about 80 of those 16,000 reads. Why could this be the case?

Not sure if this is your issue as I've never used "featureCounts", but as far as viewing alignments, I think IGV by default down samples read alignments. You can control this behavior by choosing "View" "Preferences" and going to the "Alignments" tab. If you unclick the "Downsample reads" box, then it should show all reads that meet the map quality threshold (which can also be set on this tab.)

--
Phillip

[lre1234]

For your miR-92a example, there are two homologues of this miRNA. One is on chr13 and one on chr8. The 3p arm has the same sequence for both.

>ID=MIMAT0000092; hsa-miR-92a-3p;Derives_from=MI0000093|13|+|92003615|92003636|
TATTGCACTTGTCCCGGCCTGT
--
>ID=MIMAT0000092_1; Name=hsa-miR-92a-3p;Derives_from=MI0000094|X|-|133303574|133303595|
TATTGCACTTGTCCCGGCCTGT

Not sure if this fully explains the issue you are having!

[dpryan]

At least in some annotations, that whole cluster overlaps MIR17HG, which is a "host gene" for the whole miRNA cluster. If your annotation has that then that's why you're not getting counts.

[shi]

Does hsa-miR-92a overlap with other miRNAs? If a read overlaps with more than one miRNA, featureCounts wouldnt count it.

[apredeus]

Most of your reads did not hit any miRNA listed in your provided annotation and therefore they were not counted by featureCounts (Unassigned_NoFeatures 17288648).

If your annotation only includes mature miRNA, then one possible explanation for the low counting percentage is that there are a lot of precursor miRNA in your sample.

That's not quite what I am wondering about.

I have a concrete miRNA, say, hsa-miR-92a. I see certain number of reads mapping there, about 16,000, of which about 6,000 are unique mappers. Each of those reads has a unique Illumina name, and starts on the right spot (beginning of mature 3p-miRNA).

At the same time, IGV and featureCounts recognize only about 80 of those 16,000 reads. Why could this be the case?

[shi]

Most of your reads did not hit any miRNA listed in your provided annotation and therefore they were not counted by featureCounts (Unassigned_NoFeatures 17288648).

If your annotation only includes mature miRNA, then one possible explanation for the low counting percentage is that there are a lot of precursor miRNA in your sample.

[apredeus]

yeah, sure

Status D-A1_S3_L001_R1_001.bam
Assigned 6241430
Unassigned_Ambiguity 209144
Unassigned_MultiMapping 0
Unassigned_NoFeatures 17288648
Unassigned_Unmapped 0
Unassigned_MappingQuality 0
Unassigned_FragmentLength 0
Unassigned_Chimera 0
Unassigned_Secondary 0
Unassigned_Nonjunction 0
Unassigned_Duplicate 0

[shi]

Can you show the counting summary generated by featureCounts - you can find it in the '.summary' file generated by featureCounts?

[apredeus]

Hello all,

I've recently tried to process few miRNA-seq experiments for our collaborators.

The library is quite duplicated but I guess that's to be expected with miRNA; you have about 10-fold duplicate rate for most I managed to look at with FastQC

I've got (what seems to be) the correct settings for STAR and got a fairly nice looking BAM, with some miRNAs (from miR-base) having thousands of reads according to the index. But both IGV and counting tools don't see most of those reads!

The mapping qualities are fairly good, all the names are unique (I've looked at the specific position in the BAM file manually), and about 6 thousand have NH field set to 1! So these are unique reads that clearly align in the specified location - and yet they are ignored by both IGV and most read counters (I've tried featureCounts with multimapper option on, and RSEM so far).

Help me out, I'm going insane here.