What it does
Qualimap RNA-Seq QC reports quality control metrics and bias estimations which are specific for whole transcriptome sequencing, including reads genomic origin, junction analysis, transcript coverage and 5’-3’ bias computation. As such, the tool complements the more general analysis with QualiMap BamQC, and its (optional) gene counts output can be analyzed further with QualiMap Counts QC.
Mapped reads input dataset
The dataset holding the mapped reads to carry out the analysis with. Typically, this will have been produced by a splicing-aware aligner like HISAT2 or RNA STAR.
Genome annotation data
A GTF dataset of genomic features that mapped reads should be counted for.
Counting mode
Determines whether reads should be counted individually, or whether multiple reads originating from the same sequencing template (i.e., the read and its mate in paired-end sequencing) should be counted as one.
You will usually want to choose Count fragments for paired-end data. For single-end data, choose Count reads.
Keep the per-gene counts data?
Controls whether the optional Counts output dataset should be produced, or not.
If you choose to produce this dataset, you can use:
Name to use for the counts column to specify the name of the second column in that output.
Using, for example, the name of the analyzed sample here can help you keep track of your data, especially when joining several counts datasets into a count matrix later on. In addition, Qualimap Counts QC will reuse the names of counts columns as sample names.
Read selection for counting section
Strandedness
Choose here the option that fits the strand-specificity of your sequencing library.
The Galaxy Training Material has an excellent discussion of sequencing data strandedness included in the Reference-based RNA-Seq data analysis tutorial.
Multimapping reads
Choose here how to treat reads that are mapped ambiguously to several genome locations.
Count uniquely mapped reads only excludes multi-mapping reads
Count also multimapping reads activates proportional counting of multi-mapping reads.
In this mode, each read is weighted according to the number of mapped locations. For example, a read mapped to 4 different locations will add 0.25 to the "counts" of each of the locations it maps to. The final calculated counts per feature will be converted to integer numbers.
Note: Detection of multi-mapping reads by the tool relies on the NH tag of reads in the BAM input, so make sure the aligner used to produce the dataset is configured to write this tag.
Summary Section
Reads alignment
Summarizes the mapping characteristics of the reads in the input:
total number of mapped reads
reported as left/right read mates in case of paired-end reads; excludes secondary alignments
If you accidentally selected Count fragments as the Counting mode for single-end data these and the following count of Number of aligned pairs will be zero.
total number of alignments
reports all alignment records found, including secondary alignments
number of secondary alignments
number of non-unique alignments
reports the number of alignment records with an NH tag greater than one; corresponds to the number of alignments that will have been skipped during counting when Count uniquely mapped reads only is selected
number of reads aligned to genes
number of ambiguous alignments
This is the number of mapped reads that span multiple annotated genes. Such reads are always skipped during counting.
no feature assigned
reports the number of alignments that are not overlapping any annotated feature; these may represent alignments to introns or intergenic regions, or, if the number is really high, may indicate a problem with your genome annotations
not aligned
number of reads not mapped by the aligner (but included in the BAM input)
strand specificity estimation (fwd/rev)
computed if Count reads/fragments independent of strandedness is selected; estimate of the proportion of alignments in line with forward- and reverse- strand-specificitiy of the sequencing library
Balanced proportions (i.e. ~ 0.5 forward- and ~ 0.5 reverse-strand support) can be interpreted as likely non-strand-specificity of the sequencing library, while a strand-specific library would manifest itself in a large fraction of reads supporting that specific strand-specificity.
Reads genomic origin
Lists how many alignments (absolute number/fraction) fall into
regions, or are at least
Transcript coverage profile
The profile provides ratios between mean coverage of 5’ regions, 3’ regions and whole transcripts.
5’ bias
the ratio of coverage median of 5’ regions (defined as the first 100 nts) to whole transcripts
3' bias
the ratio of coverage median of 3’ regions (defined as the last 100 nts) to whole transcripts
5’-3’ bias
the ratio of 5' bias to 3' bias.
Junction analysis
Lists the total number of reads with splice junctions and the relative frequency of the (up to) 10 most frequent junction sequences.
Plots
Reads Genomic Origin
A pie chart showing how many read alignments fall into exonic, intronic and intergenic regions.
Coverage Profile Along Genes (Total)
This plot shows the mean coverage profile of all genes with non-zero overall coverage.
Coverage Profile Along Genes (Low)
The plot shows the mean coverage profile of the 500 genes with the lowest, but non-zero overall coverage.
Coverage Profile Along Genes (High)
The plot shows the mean coverage profile of the 500 genes with the highest overall coverage.
Coverage Histogram (0-50x)
Coverage of genes from 0 to 50x. Genes with >50x coverage are added to the 50x bin.
Junction Analysis
This pie chart shows an analysis of the splice junctions observed in the alignments. It consists of three categories:
Known
observed splice junctions both sides of which are in line with the genome annotation data
Partly known
observed splice junctions for which only one junction side can be deduced from the genome annotation data
Novel
observed splice junctions not predicted on either side by the genome annotation data
This is a Collection of 4 individual datasets.
Of these, the rnaseq_qc_results dataset provides a plain-text version of the HTML report Summary section.
The other 3 datasets hold the tabular raw data underlying the three coverage profile plots in the HTML Report.
Optional. This is a 2-column tabular dataset of read or fragment counts (depending on the chosen Counting mode) per annotated gene. The first column lists the gene identifiers found in the Genome annotation data, the second the associated counts.
This dataset represents valid (single-sample) input for the QualiMap Counts QC tool.