output.md

Output

This document describes the output produced by the pipeline. Most of the plots are taken from the MultiQC report, which summarises results at the end of the pipeline.

Pipeline overview

The pipeline is built using Nextflow and processes whole-exome sequencing (WES), or whole-genome sequencing (WGS) data following the steps presented in the main README file.

Briefly, the workflow runs several quality controls from the raw and aligned data in order to validate both frozen and FFPE samples. Then, several tools can be run in order to detect germline single nucleotide variants (SNVs) with haplotypecaller, somatic SNVs with mutect2, structural variants (SVs) with MANTA or copy number variants (CNV) with ASCAT and FACETS. Finaly, complementary analysis based on variant calling can also be performed. Annotation can be done with SnpEff and SnpSift. Tumor Mutational Burden is assessed with pyTMB tool and Microsatellite Instability is assesed with MSIsensor-pro.

The directories listed below will be created in the output directory after the pipeline has finished.

FastQC

FastQC gives general quality metrics about your reads. It provides information about the quality score distribution across your reads, the per base sequence content (%T/A/G/C). You get information about adapter contamination and other overrepresented sequences.

For further reading and documentation see the FastQC help.

NB: The FastQC plots displayed in the MultiQC report shows the input reads. In theory, they should be already trimmed for adapter sequence and potentially regions with low quality. For details about reads trimming, see the raw_qc pipeline.

Output directory: fastqc

• [SAMPLE]_fastqc.html
  • FastQC report, containing quality metrics for your untrimmed raw fastq files
• zips/[SAMPLE]_fastqc.zip
  • zip file containing the FastQC report, tab-delimited data file and plot images

Preprocessing

Alignment

Raw reads are aligned on the reference genome by default with BWA-mem. The mapping statistics (Total Reads, Aligned Reads, High-Quality Alignment, Low-Quality Alignment) are also presented in the main summary table.
Note that if multiple sequencing lanes from the same samples (same sampleID, sampleName) are specified, the bam files are merged just after BWA-mem.

Output directory: preprocessing/bams/bwa/

• [SAMPLE].bam and [SAMPLE].bam.bai
  • Aligned reads with BAM index

The mapping statistics are presented in the MultiQC report as follows.
In general, we expect more than 80% of aligned reads. Samples with less than 50% of mapped reads should be further investigated, and check for adapter content, contamination, etc.

NB: Note that by default, these mapping files are not saved. Use --saveAlignedIntermediates to save them.

Duplicates

Sambamba is used to mark the duplicates. The results are presented in the General Metrics table.

Output directory: preprocessing/bams/markDuplicates

• [SAMPLE].md.bam

  • Aligned reads marked for duplicates

• stats/[SAMPLE].md.flagstats

  • Number of alignments for each FLAG type

  NB: Note that by default, these mapping files are not saved. Use --saveAlignedIntermediates to save them.

Reads on target

In the context of WES analysis, the aligned reads are intersected with their targets, defined with the --targetBed parameter. The percentage of reads on targets are presented in the General Metrics table.

Output directory: preprocessing/bams/onTarget

• [SAMPLE].onTarget.bam

  • Aligned reads restricted to the genomic targets.

• stats/[SAMPLE].onTarget.flagstat

  • Number of alignments for each FLAG type

  NB: Note that by default, these mapping files are not saved. Use --saveAlignedIntermediates to save them.

Filtering

Aligned reads are then filtered-out in order to remove non informative reads for the downstream analysis.

The aligned reads can be filtered out as follow :

• mapq : discard reads aligned with a mapping quality lower than --mapQual
• --keepDups false : discard reads flagged as duplicates
• --keepSingleton false : discard reads for which the paired mate is not aligned
• --keepMultiHits : discard reads aligned several times on the genome

By default the filters are defined to remove low mapq, dupicated, singleton and multi hit reads.
The fraction of remaining reads after filtering is also presented in the General Metrics table.

Output directory: preprocessing/bams/filtering/

• [SAMPLE].filtered.bam and [SAMPLE].filtered.bam.bai
  • Aligned and filtered reads with BAM index
• [SAMPLE].filtered.idxstats
  • Alignment summary statistics
• [SAMPLE].filtered.flagstats
  • Number of alignments for each FLAG type

Quality controls

From the filtered and aligned reads files, the pipeline then runs several quality control steps presented below.

Sequencing complexity

The Preseq package is aimed at predicting and estimating the complexity of a genomic sequencing library, equivalent to predicting and estimating the number of redundant reads from a given sequencing depth and how many will be expected from additional sequencing using an initial sequencing experiment. The estimates can then be used to examine the utility of further sequencing, optimize the sequencing depth, or to screen multiple libraries to avoid low complexity samples. The dashed line shows a perfectly complex library where total reads = unique reads. Note that these are predictive numbers only, not absolute. The MultiQC plot can sometimes give extreme sequencing depth on the X axis - click and drag from the left side of the plot to zoom in on more realistic numbers.

Output directory: preprocessing/metrics/preseq

• [SAMPLE]_ccurve.txt

  • Preseq expected future yield file.

  

Fragment length

The fragment length is calculated from paired-end reads as the distance between the two mates with picard. The mean value is presented in the General Metrics table and the distribution is presented by MultiQC as follow :

Output directory: preprocessing/metrics/fragSize

• [SAMPLE]_insert_size_metrics.txt

  • Fragment size values reported by picard

• [SAMPLE]_insert_size_hist.pdf

  • Graphical representation

  

Sequencing depth

The mean sequencing depth and the percentage of the genome (or targets) covered at soem threshold (X) are calculated with mosdepth.
The coverage at 30X, 50X (hidden column) and 100X are available in the General Metrics table.

In addition, the same analysis is repeated for exonic regions only. In the context of WES analysis, only the exonic regions overlapping with the targets are used.
The results are presented in the 'Genes Coverage' section of the MultiQC report.

Output directory: preprocessing/metrics/depth

• *global* files are the mosdepth outputs for stantard coverage on the genome
• *regions* files are the mosdepth outputs for the gene coverage

WGS metrics

The picard collectWgsMetrics tool is run to collect some additional statistics on reads mapping. Among them, the fraction of bases covered by both R1 and R2 mates are available in the General Metrics table.
In the case of FFPE samples for which the fragment size is usually smaller, this metric can help adjusting the sequencing length. In addition, overlapping read pairs can sometimes be an issue for downstream analysis, and a reads trimming (or merge) can be an interesting option.

Identity monitoring

In order to check the association between pairs of normal/tumor samples, a list of common SNPs (--polym) is used to cluster all the samples.
The results are displayed as a dendrogram.

Output directory: Identito

• [SAMPLE].matrix.tsv

  • results of the SNPs calling for the list of SNPs

• clustering_plot_identito.csv

  • Identito dendrogram

  

SNVs calling

GATK Preprocessing

The current workflow follows the GATK good practices with base recalibration.
This step is usally recommanded to detects systematic errors in the data, but can be skipped with the option --skipBQSR.
These files are used as inputs of all germline and somatic SNVs calling.

Output directory: preprocessing/bams/bqsr

• [SAMPLE].recal.bam and [SAMPLE].recal.bam.bai
  • Aligned data after base recalibration with BAM index

Germline variants

Germline variants are then called using haplotypecaller following good practices (HaplotypeCaller, GenotypeGVCFs). The number of detected variants are presented as a table in the MultiQC report.

Output directory: HaplotypeCaller

• [SAMPLE]_HaplotypeCaller.vcf.gz and [SAMPLE]_HaplotypeCaller.vcf.gz.tbi
  • vcf file with the variants detected by HaplotypeCaller with Tabix index

Somatic mutations

The somatic mutations calling requires pairs of normal/tumor samples defined in the design file.
The mutect2 tool is used to call somatic variants following the GATK good practices (Mutect2, MergeMutectStats, GetPileupSummaries, GatherPileupSummaries, CalculateContamination, LearnReadOrientationModel, FilterMutectCall).

Output directory: Mutect2

• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_Mutect2_unfiltered.vcf.gz and [TUMORSAMPLE]_vs_[NORMALSAMPLE]_Mutect2_unfiltered.vcf.gz.tbi
  • Mutect2 somatic variants before filtering with Tabix index
• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_Mutect2_filtered.vcf.gz and [TUMORSAMPLE]_vs_[NORMALSAMPLE]_Mutect2_filtered.vcf.gz.tbi
  • Mutect2 somatic variants after filtering with Tabix index
• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_Mutect2_filtered_pass_norm.vcf.gz and [TUMORSAMPLE]_vs_[NORMALSAMPLE]_Mutect2_filtered_pass_norm.vcf.gz.tbi
  • Mutect2 somatic variants after filtering for PASS variants only and normalization with bcftools norm and with Tabix index

Transition/Transversion

For each filtered vcf files, the current workflow calculate the number of transition (A>G,T>C,C>T,G>A), transversions (A>C,T>G,C>A,G>T,A>T,T>A,C>G,G>C) and short insertions/delations (indels). The results are available as table and presented in MultiQC.

Output directory: HaplotypeCaller/[SAMPLE]/tstv

• [SAMPLE]_filtered.vcf.Mutect2.table.tsv
  • Number of bases substitution and indels for each type

Output directory: Mutect2/[SAMPLE]/tstv

• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_filtered.vcf.Mutect2.table.tsv

  • Number of bases substitution and indels for each type

  

Variants annotation

Each filtered VCF file is then annotated using snpeff and SnpSift. All annotated vcf files are saved and the following summary metrics are displayed in MultiQC.

Output directory: snpEff/[SAMPLE]

• [SAMPLE]_snpeff.ann.vcf.gz
  • annotated variants with SnpEff
• [SAMPLE]_dbNSFP.vcf.gz
  • annotated variants with SnpSift and DBNSFP database
• [SAMPLE]_GnomAD.vcf.gz
  • annotated variants with SnpSift and Gnomad database
• [SAMPLE]_CancerHotspots.vcf.gz
  • annotated variants with SnpSift and CancerHotspots database
• [SAMPLE]_ICGC.vcf.gz
  • annotated variants with SnpSift and ICGC database
• [SAMPLE]_COSMIC.vcf.gz
  • annotated variants with SnpSift and COSMIC database

CNVs calling

CNVs calling can be run using ASCAT and FACETS. Both tools require pairs of tumor/normal samples. ASCAT and Facets are two software for performing allele-specific copy number analysis of tumor samples and for estimating tumor ploidy and purity (normal contamination). They infer tumor purity and ploidy and calculates allele-specific copy number profiles. Both tools provide several images and tables as output.

Output directory: Facets

• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_cnv_amp_del.txt
  • list of amplification and deletion of potential interest.
• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_cnv.pdf
  • CNVs plot
• [TUMORSAMPLE]_vs_[NORMALSAMPLE]_cnv_ploidy_cellularity.txt
  • table with cellularity and ploidy for this sample.

Output directory: ASCAT

• [TUMORSAMPLE].BAF and [NORMALSAMPLE].BAF
  • file with beta allele frequencies generated by AlleleCount
• [TUMORSAMPLE].LogR and [NORMALSAMPLE].LogR
  • file with total copy number on a logarithmic scale generated by AlleleCount
• [TUMORSAMPLE].ASCATprofile.png
  • Image with information about ASCAT profile
• [TUMORSAMPLE].ASPCF.png
  • Image with information about ASPCF
• [TUMORSAMPLE].rawprofile.png
  • Image with information about raw profile
• [TUMORSAMPLE].sunrise.png
  • Image with information about sunrise
• [TUMORSAMPLE].tumour.png
  • Image with information about tumor
• [TUMORSAMPLE].cnvs.txt
  • file with information about CNVS
• [TUMORSAMPLE].LogR.PCFed.txt
  • file with information about LogR
• [TUMORSAMPLE].BAF.PCFed.txt
  • file with information about BAF
• [TUMORSAMPLE].purityploidy.txt
  • file with information about purity ploidy

The text file [TUMORSAMPLE].cnvs.txt countains predictions about copy number state for all the segments. The output is a tab delimited text file with the following columns:

• chr: chromosome number
• startpos: start position of the segment
• endpos: end position of the segment
• nMajor: number of copies of one of the allels (for example the chromosome inherited from the father)
• nMinor: number of copies of the other allele (for example the chromosome inherited of the mother)

SVs calling

Structural variants and indels are called using MANTA with matched control. It is optimized for analysis of germline variation in small sets of individuals and somatic variation in tumor/normal sample pairs.

For all samples :

Output directory: Manta

• Manta_[SAMPLE].candidateSmallIndels.vcf.gz and Manta_[SAMPLE].candidateSmallIndels.vcf.gz.tbi
  • VCF with Tabix index
• Manta_[SAMPLE].candidateSV.vcf.gz and Manta_[SAMPLE].candidateSV.vcf.gz.tbi
  • VCF with Tabix index

For Normal sample only:

• Manta_[NORMALSAMPLE].diploidSV.vcf.gz and Manta_[NORMALSAMPLE].diploidSV.vcf.gz.tbi
  • VCF with Tabix index

For a Tumor sample only:

• Manta_[TUMORSAMPLE].tumorSV.vcf.gz and Manta_[TUMORSAMPLE].tumorSV.vcf.gz.tbi
  • VCF with Tabix index

For Tumor/Normal pair :

Output directory: Manta

• Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].candidateSmallIndels.vcf.gz and Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].candidateSmallIndels.vcf.gz.tbi
  • VCF with Tabix index
• Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].candidateSV.vcf.gz and Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].candidateSV.vcf.gz.tbi
  • VCF with Tabix index
• Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].diploidSV.vcf.gz and Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].diploidSV.vcf.gz.tbi
  • VCF with Tabix index
• Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].somaticSV.vcf.gz and Manta_[TUMORSAMPLE]_vs_[NORMALSAMPLE].somaticSV.vcf.gz.tbi
  • VCF with Tabix index

Complementary analysis:

Microsatellite instability

Mictosatellite instability is assesed with MSIsensor-pro.

Output directory: MSI/[SAMPLE]

• [SAMPLE]
  • The final report contain all detected microsatellites , the unstable(somatic) microsatellites and the MSI score.
• [SAMPLE]_dis
• [SAMPLE]_germline
• [SAMPLE]_somatic

TMB

Tumor Mutational Burden is assesed with pyTMB?

Output directory: MSI/[SAMPLE]

• [SAMPLE]_tmb.txt
  • results file

Table Report

This final step transforms the fully annotated vcf into a tabulated file to ease the analysis of the results.

Output directory: tableReport/[SAMPLE]

• [SAMPLE].tsv
  • results file

MultiQC

MultiQC is a visualisation tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in within the report data directory.

The pipeline has special steps which allow the software versions used to be reported in the MultiQC output for future traceability.

Output directory: results/multiqc

• Project_multiqc_report.html
  • MultiQC report - a standalone HTML file that can be viewed in your web browser
• Project_multiqc_data/
  • Directory containing parsed statistics from the different tools used in the pipeline

For more information about how to use MultiQC reports, see http://multiqc.info. See the file 'test/multiqc_report.html' for an example on the test dataset.