A modular tool to aggregate results from bioinformatics analyses across many samples into a single report.
This report has been generated by the nf-core/rnaseq analysis pipeline. For information about how to interpret these results, please see the documentation.
Report
generated on 2023-09-05, 04:47 CEST
based on data in:
/data/ceph/hdd/project/node_05/ghga-rnaseq-harmonization/work/90/3aeff1740cae42c301e526c1994301
General Statistics
Showing 1998 samples.
DupRadar
provides duplication rate quality control for RNA-Seq datasets. Highly expressed genes can be expected to have a lot of duplicate reads, but high numbers of duplicates at low read counts can indicate low library complexity with technical duplication. This plot shows the general linear models - a summary of the gene duplication distributions. .
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
STAR_RSEM DESeq2 PCA plot
PCA plot between samples in the experiment. These values are calculated using DESeq2 in the deseq2_qc.r
script.
STAR_RSEM DESeq2 sample similarity
is generated from clustering by Euclidean distances between DESeq2 rlog values for each sample in the deseq2_qc.r
script.
Biotype Counts
shows reads overlapping genomic features of different biotypes, counted by featureCounts.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Picard
Picard is a set of Java command line tools for manipulating high-throughput sequencing data.
Mark Duplicates
Number of reads, categorised by duplication state. Pair counts are doubled - see help text for details.
The table in the Picard metrics file contains some columns referring read pairs and some referring to single reads.
To make the numbers in this plot sum correctly, values referring to pairs are doubled according to the scheme below:
READS_IN_DUPLICATE_PAIRS = 2 * READ_PAIR_DUPLICATES
READS_IN_UNIQUE_PAIRS = 2 * (READ_PAIRS_EXAMINED - READ_PAIR_DUPLICATES)
READS_IN_UNIQUE_UNPAIRED = UNPAIRED_READS_EXAMINED - UNPAIRED_READ_DUPLICATES
READS_IN_DUPLICATE_PAIRS_OPTICAL = 2 * READ_PAIR_OPTICAL_DUPLICATES
READS_IN_DUPLICATE_PAIRS_NONOPTICAL = READS_IN_DUPLICATE_PAIRS - READS_IN_DUPLICATE_PAIRS_OPTICAL
READS_IN_DUPLICATE_UNPAIRED = UNPAIRED_READ_DUPLICATES
READS_UNMAPPED = UNMAPPED_READS
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
QualiMap
QualiMap is a platform-independent application to facilitate the quality control of alignment sequencing data and its derivatives like feature counts.DOI: 10.1093/bioinformatics/btv566; 10.1093/bioinformatics/bts503.
Genomic origin of reads
Classification of mapped reads as originating in exonic, intronic or intergenic regions. These can be displayed as either the number or percentage of mapped reads.
There are currently three main approaches to map reads to transcripts in an RNA-seq experiment: mapping reads to a reference genome to identify expressed transcripts that are annotated (and discover those that are unknown), mapping reads to a reference transcriptome, and de novo assembly of transcript sequences (Conesa et al. 2016).
For RNA-seq QC analysis, QualiMap can be used to assess alignments produced by the first of these approaches. For input, it requires a GTF annotation file along with a reference genome, which can be used to reconstruct the exon structure of known transcripts. This allows mapped reads to be grouped by whether they originate in an exonic region (for QualiMap, this may include 5′ and 3′ UTR regions as well as protein-coding exons), an intron, or an intergenic region (see the Qualimap 2 documentation).
The inferred genomic origins of RNA-seq reads are presented here as a bar graph showing either the number or percentage of mapped reads in each read dataset that have been assigned to each type of genomic region. This graph can be used to assess the proportion of useful reads in an RNA-seq experiment. That proportion can be reduced by the presence of intron sequences, especially if depletion of ribosomal RNA was used during sample preparation (Sims et al. 2014). It can also be reduced by off-target transcripts, which are detected in greater numbers at the sequencing depths needed to detect poorly-expressed transcripts (Tarazona et al. 2011).
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Gene Coverage Profile
Mean distribution of coverage depth across the length of all mapped transcripts.
There are currently three main approaches to map reads to transcripts in an RNA-seq experiment: mapping reads to a reference genome to identify expressed transcripts that are annotated (and discover those that are unknown), mapping reads to a reference transcriptome, and de novo assembly of transcript sequences (Conesa et al. 2016).
For RNA-seq QC analysis, QualiMap can be used to assess alignments produced by the first of these approaches. For input, it requires a GTF annotation file along with a reference genome, which can be used to reconstruct the exon structure of known transcripts. QualiMap uses this information to calculate the depth of coverage along the length of each annotated transcript. For a set of reads mapped to a transcript, the depth of coverage at a given base position is the number of high-quality reads that map to the transcript at that position (Sims et al. 2014).
QualiMap calculates coverage depth at every base position of each annotated transcript. To enable meaningful comparison between transcripts, base positions are rescaled to relative positions expressed as percentage distance along each transcript (0%, 1%, …, 99%). For the set of transcripts with at least one mapped read, QualiMap plots the cumulative mapped-read depth (y-axis) at each relative transcript position (x-axis). This plot shows the gene coverage profile across all mapped transcripts for each read dataset. It provides a visual way to assess positional biases, such as an accumulation of mapped reads at the 3′ end of transcripts, which may indicate poor RNA quality in the original sample (Conesa et al. 2016).
The Normalised plot is calculated by MultiQC to enable comparison of samples with varying sequencing depth. The cumulative mapped-read depth at each position across the averaged transcript position are divided by the total for that sample across the entire averaged transcript.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Rsem
Rsem RSEM (RNA-Seq by Expectation-Maximization) is a software package forestimating gene and isoform expression levels from RNA-Seq data.DOI: 10.1186/1471-2105-12-323.
Mapped Reads
A breakdown of how all reads were aligned for each sample.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Multimapping rates
A frequency histogram showing how many reads were aligned to n
reference regions.
In an ideal world, every sequence reads would align uniquely to a single location in the reference. However, due to factors such as repeititve sequences, short reads and sequencing errors, reads can be align to the reference 0, 1 or more times. This plot shows the frequency of each factor of multimapping. Good samples should have the majority of reads aligning once.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
RSeQC
RSeQC package provides a number of useful modules that can comprehensively evaluate high throughput RNA-seq data.DOI: 10.1093/bioinformatics/bts356.
Read Distribution
Read Distribution calculates how mapped reads are distributed over genome features.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Inner Distance
Inner Distance calculates the inner distance (or insert size) between two paired RNA reads. Note that this can be negative if fragments overlap.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Read Duplication
read_duplication.py calculates how many alignment positions have a certain number of exact duplicates. Note - plot truncated at 500 occurrences and binned.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Junction Annotation
Junction annotation compares detected splice junctions to a reference gene model. An RNA read can be spliced 2 or more times, each time is called a splicing event.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Junction Saturation
Junction Saturation counts the number of known splicing junctions that are observed in each dataset. If sequencing depth is sufficient, all (annotated) splice junctions should be rediscovered, resulting in a curve that reaches a plateau. Missing low abundance splice junctions can affect downstream analysis.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Infer experiment
Infer experiment counts the percentage of reads and read pairs that match the strandedness of overlapping transcripts. It can be used to infer whether RNA-seq library preps are stranded (sense or antisense).
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Bam Stat
All numbers reported in millions.
Samtools
Samtools is a suite of programs for interacting with high-throughput sequencing data.DOI: 10.1093/bioinformatics/btp352.
Percent Mapped
Alignment metrics from samtools stats
; mapped vs. unmapped reads.
For a set of samples that have come from the same multiplexed library, similar numbers of reads for each sample are expected. Large differences in numbers might indicate issues during the library preparation process. Whilst large differences in read numbers may be controlled for in downstream processings (e.g. read count normalisation), you may wish to consider whether the read depths achieved have fallen below recommended levels depending on the applications.
Low alignment rates could indicate contamination of samples (e.g. adapter sequences), low sequencing quality or other artefacts. These can be further investigated in the sequence level QC (e.g. from FastQC).
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Alignment metrics
This module parses the output from samtools stats
. All numbers in millions.
Samtools Flagstat
This module parses the output from samtools flagstat
. All numbers in millions.
XY counts
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Mapped reads per contig
The samtools idxstats
tool counts the number of mapped reads per chromosome / contig. Chromosomes with < 0.1% of the total aligned reads are omitted from this plot.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
FastQC (raw)
FastQC (raw) This section of the report shows FastQC results before adapter trimming.
Sequence Counts
Sequence counts for each sample. Duplicate read counts are an estimate only.
This plot show the total number of reads, broken down into unique and duplicate if possible (only more recent versions of FastQC give duplicate info).
You can read more about duplicate calculation in the FastQC documentation. A small part has been copied here for convenience:
Only sequences which first appear in the first 100,000 sequences in each file are analysed. This should be enough to get a good impression for the duplication levels in the whole file. Each sequence is tracked to the end of the file to give a representative count of the overall duplication level.
The duplication detection requires an exact sequence match over the whole length of the sequence. Any reads over 75bp in length are truncated to 50bp for this analysis.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Sequence Quality Histograms
The mean quality value across each base position in the read.
To enable multiple samples to be plotted on the same graph, only the mean quality scores are plotted (unlike the box plots seen in FastQC reports).
Taken from the FastQC help:
The y-axis on the graph shows the quality scores. The higher the score, the better the base call. The background of the graph divides the y axis into very good quality calls (green), calls of reasonable quality (orange), and calls of poor quality (red). The quality of calls on most platforms will degrade as the run progresses, so it is common to see base calls falling into the orange area towards the end of a read.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Per Sequence Quality Scores
The number of reads with average quality scores. Shows if a subset of reads has poor quality.
From the FastQC help:
The per sequence quality score report allows you to see if a subset of your sequences have universally low quality values. It is often the case that a subset of sequences will have universally poor quality, however these should represent only a small percentage of the total sequences.
Flat image plot. Toolbox functions such as highlighting / hiding samples will not work (see the docs).
Per Base Sequence Content
The proportion of each base position for which each of the four normal DNA bases has been called.
To enable multiple samples to be shown in a single plot, the base composition data is shown as a heatmap. The colours represent the balance between the four bases: an even distribution should give an even muddy brown colour. Hover over the plot to see the percentage of the four bases under the cursor.
To see the data as a line plot, as in the original FastQC graph, click on a sample track.
From the FastQC help:
Per Base Sequence Content plots out the proportion of each base position in a file for which each of the four normal DNA bases has been called.
In a random library you would expect that there would be little to no difference between the different bases of a sequence run, so the lines in this plot should run parallel with each other. The relative amount of each base should reflect the overall amount of these bases in your genome, but in any case they should not be hugely imbalanced from each other.
It's worth noting that some types of library will always produce biased sequence composition, normally at the start of the read. Libraries produced by priming using random hexamers (including nearly all RNA-Seq libraries) and those which were fragmented using transposases inherit an intrinsic bias in the positions at which reads start. This bias does not concern an absolute sequence, but instead provides enrichement of a number of different K-mers at the 5' end of the reads. Whilst this is a true technical bias, it isn't something which can be corrected by trimming and in most cases doesn't seem to adversely affect the downstream analysis.