miRNA-seq quality control is frequently restricted to calculating
PhredScores and filtering of low complexity reads since miRNA specific quality measures cannot be easily assessed.
mirnaQC provides several quality features that can help researchers identify issues in their samples. These features are
provided as absolute values and ranked with a percentile calculated from a corpus of more than 36,000 samples.
Please note that mirnaQC was developed and tested using Google Chrome and Mozilla Firefox. Using other (older) browsers
will likely result in some features not working properly. Please contact us to try to solve them.
mirnaQC was developed by the Computational Epigenomics Group at University of Granada
Contact us
If you have any issues or doubts using mirnaQC please contact hackenberg at go.ugr.es
Other miRNA tools from our lab
If you use and like mirnaQC consider other tools from our lab:
We have prepared two step-by-step tutorials to help users launch mirnaQC jobs and interpret the output page:
Values and percentiles
For each of the quality parameters calculated by mirnaQC you can display the raw value or the percentile of this
value in the whole of Sequence Reads Archive.
To calculate these distributions, we analysed more than 36,000 miRNA-seq samples from 25 different species.
Frequently Asked Questions (FAQs)
Q: Which percentiles indicate good quality?
A: For some attributes, low values are good (adapter dimers or percentage of ribosomal RNA fragments) while
for others higher numbers are better (microRNA yield, percentage of reads in the analysis, Phred Scores).
For each attribute and sample we provide the percentile the value has in the background distribution
(the reference corpus). Therefore, for adapter-dimers, lower quartile values are good, while for microRNA
yield upper quartile values would be desired. However, in order to keep the color code coherent the
‘good quartile’ will always be displayed in green while ‘the bad quartile’ will appear in red.
Q: How is the heatmap generated?
In order to generate a unique colour scheme in the heatmap, the percentiles need to be transformed
so that the lower quartile indicates good performance while the upper quartile indicates bad
results (see also What are the good percentiles? above for more information).
Q: Is SOLID data supported?
A: mirnaQC was conceptually designed for Illumina (nucleotide space) data. That implies that
the pipeline can process SOLID data but many attributes will be meaningless. This is because to
analyse SOLID data, colour space reads are mapped to the genome first in order to perform the
conversion to nucleotide space. This means that the percentage of mapped reads or even adapter
dimers cannot be assessed. Other attributes like the number of miRNAs, percentage of short reads,
library complexity can be correctly analysed and compared.
PCA tool info
This tool will display scatter-plot of a PCA calculated from the expression values of your data. You can then explore the different
Principal Components using the selectors.
You can also change which quality parameter is used to color samples. This may help you spot variables that are driving
the expression in your samples.
You can mouse over any point in the plot to see a sample percentile and value. Group information is also available if provided
and each group will be represented using different symbols (square, circle, cross, etc).