Skip to content

Mutational Signatures

Authors Scott Newman, Michael Macias
Publication Mutational Signatures employs MutationalPatterns: "MutationalPatterns: comprehensive genome-wide analysis of mutational processes."
Technical Support Contact Us

Mutational Signatures finds and quantifies COSMIC mutational signatures across samples. This is done by finding the optimal non-negative linear combination of mutation signatures to reconstruct a mutation matrix. It builds the initial mutation matrix from multiple single-sample VCFs and, by default, fits it to mutational signatures from COSMIC. Mutational Signatures employs MutationalPatterns (Blokzijl, et al. (2018)) to achieve this.

Mutational Signatures supports both hg19 (GRCh37) and hg38 (GRCh38).

Overview

Inputs

Name Type Description Example
VCF(s) Array of files List of VCF inputs. Can be single-sample or multi-sample and uncompressed or gzipped. [*.vcf, *.vcf.gz]
Sample sheet File Tab-delimited file (no headers) with sample ID and tag pairs [optional] *.txt
Genome build String Genome build used as reference. Can be either "GRCh37" or "GRCh38". [default: "GRCh38"] GRCh38
Minimum mutation burden Integer Minimum number of somatic SNVs a sample must have to be considered for analysis [default: 9] 15
Minimum signature contribution Integer Minimum number of mutations attributable to a single signature [default: 9] 100
Output prefix String Prefix to append to output filenames [optional] mtsg
Disabled VCF column Integer VCF column (starting from sample names, zero-based) to ignore when reading VCFS [optional] 1

Outputs

Name Type Description
Raw signatures File Tab-delimited file of the raw results with sample contributions for each signature
Signatures visualization File HTML file for interactive plotting
Sample sheet File Tab-delimited file (no headers) with sample ID and tag pairs

Process

Mutational Signatures runs four steps using subcommands of mtsg.

  1. Split VCFs (single or multi-sample) to multiple single-sample VCFs.
  2. If not given, generate a sample sheet from the directory of single-sample VCFs.
  3. Build a mutation matrix and reconstruct/fit it using COSMIC mutation signatures.
  4. Create a visualization file using the fitted signatures.

Getting started

After logging in, click the "Start" button on the Mutational Signatures tool page. This creates a new DNAnexus project and imports the tool.

With subsequent runs, the sidebar shows "Launch Tool", meaning the project with the tool already exists. Click "Launch Tool" to start a new analysis.

Input configuration

Mutational Signatures only requires VCFs as inputs. This can be a single multi-sample VCF, multiple single-sample VCFs, or a combination of both. All other inputs are optional.

Input files can be uploaded via the data transfer application or command line.

VCF(s)

VCF(s) is a list of VCF inputs. The inputs can be single-sample or multi-sample and uncompressed or gzipped. Sample names are taken from the VCF header.

When using multi-sample VCFs, empty cells/absent variant calls must be denoted with .:..

gVCFs are not supported.

Sample sheet

Sample sheet is a tab-delimited file (no headers) with two columns: the sample ID and a tag. The tag is an arbitrary identifier used to group the samples, typically a disease abbreviation or tissue of origin.

If not given, a sample sheet will be generated automatically.

Example
SJACT001_D ACT
SJACT002_D ACT
SJBALL063_D BALL
SJHGG017_D HGG

Output prefix

Output prefix is the prefix to append to the output filenames. By default, if a single input VCF is given, its basename is used as the output prefix. If multiple input VCFs are given, a default "mtsg" prefix is used. This behavior can be overridden by a user-defined prefix.

Example
VCF(s) Prefix Output filename for raw signatures
[pcgp.b38.refseq.goodbad.vcf] pcgp.b38.refseq.goodbad pcgp.b38.refseq.goodbad.signatures.txt
[SJOS013_D.vcf, SJRHB007_D.vcf] mtsg mtsg.signatures.txt

Disabled VCF column

Disabled VCF column is the column index to ignore when reading VCFs. This is useful when the inputs are tumor-normal VCFs, and one column should be ignored. Otherwise, the results would likely be duplicated.

The argument is a zero-based index relative to the sample names in the header of the VCF. For example, in a VCF with samples SJEPD003_D and SJEPD003_G, the germline sample (SJEPD003_G) can be discarded by setting the disabled VCF column to 1.

Example
0 1
#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT SJEPD003_D SJEPD003_G

Uploading data

Mutational Signatures requires at least one VCF and an optional sample sheet to be uploaded. These files can be uploaded via the data transfer application or command line.

Analysis of results

Upon a successful run of Mutational Signatures, three files are saved to the results directory: raw signature contributions, a visualization file, and a sample sheet.

Interpreting results

Raw signatures

Raw signatures is a tab-delimited file of the raw results with sample contributions for each signature. Column 1 is the sample name, columns 2-(N-1) are the COSMIC signatures contribution counts, and column N is the group tag, where N is the total number of columns. The number of columns is variable since if the signature has no contributions for all samples, it is completely omitted.

Note that the last column tissue is a misnomer. It aligns to the arbitrary tag given in the sample sheet.

Example
Signature.1 Signature.2 Signature.30 tissue
SJACT001_D 1.71758029 131.033723 18.6910151 ACT
SJAMLM7005_D 51.9627312 7.10850351 0 AMLM7

Signatures visualization

Signatures visualization is an HTML file that can be used for interactive plotting.

When opened in a web browser, a set of controls allows plotting various stacked bar charts: total contributions by signature, total contributions by tag, and total contributions by sample per tag. The total contributions can be stacked as absolute values or as a percentage of the total.

Sample sheet

When no sample sheet is given as an input, one is generated automatically, but it is not guaranteed the derived tags will be of any use. This generated sample sheet is given as an output in the case the tags need to be manually edited, and the job is resubmitted with it as an input.

When a sample sheet is given as an input, the sample sheet output is a copy of the input.

See also the description for the input sample sheet.

Troubleshooting

To troubleshoot a failed run of Mutational Signatures, check the job log for details.

Wrong genome build

If the "Building mutation matrix" step during run fails, it is likely that the selected genome build does not match the input VCF(s). Rerun the job with a matching genome build.

Example

R: Building mutation matrix from 6 VCFs
R: Error in mut_matrix(vcf_list = filtered_vcfs, ref_genome = ref_genome) :
R:   Error in .Call2("solve_user_SEW", refwidths, start, end, width, translate.negative.coord,  :
R:   solving row 526: 'allow.nonnarrowing' is FALSE and the supplied start (79440206) is > refwidth + 1

References

  • Blokzijl F, Janssen R, van Boxtel R, Cuppen E (2018). "MutationalPatterns: comprehensive genome-wide analysis of mutational processes." Genome Medicine. doi: 10.1186/s13073-018-0539-0. PMID: 29695279.