Awesome

SigProfilerAssignmentR

An R wrapper for running the SigProfilerAssignment framework.

SigProfilerAssignment enables assignment of previously known mutational signatures to individual samples and individual somatic mutations. The tool refits different types of reference mutational signatures, including COSMIC signatures, as well as custom signature databases. Refitting of known mutational signatures is a numerical optimization approach that not only identifies the set of operative mutational signatures in a particular sample, but also quantifies the number of mutations assigned to each signature found in that sample. SigProfilerAssignment makes use of SigProfilerMatrixGenerator and SigProfilerPlotting, seamlessly integrating with other SigProfiler tools.

For users that prefer working in a Python environment, please check: https://github.com/AlexandrovLab/SigProfilerAssignment. Detailed documentation can be found at: https://osf.io/mz79v/wiki/home/.

Table of contents

• Installation
• Running
  • Main Parameters
  • Signature Subgroups
• Examples
• Copyright
• Contact Information

<a name="installation"></a> Installation

PREREQUISITES

devtools (R)

>> install.packages("devtools")

reticulate* (R)

>> install.packages("reticulate")  

*Reticulate has a known bug of preventing python print statements from flushing to standard out. As a result, some of the typical progress messages are delayed.

QUICK START GUIDE

1. First, install the python package using pip. The R wrapper still requires the python package:

$ pip install SigProfilerAssignment

2. Open an R session and ensure that your R interpreter recognizes the path to your python installation:

$ R
>> library(reticulate)
>> use_python("path_to_your_python")
>> py_config()
python:         /anaconda3/bin/python
libpython:      /anaconda3/lib/libpython3.6m.dylib
pythonhome:     /anaconda3:/anaconda3
version:        3.6.5 |Anaconda, Inc.| (default, Apr 26 2018, 08:42:37)  [GCC 4.2.1 Compatible Clang 4.0.1 (tags/RELEASE_401/final)]
numpy:          /anaconda3/lib/python3.6/site-packages/numpy
numpy_version:  1.16.1

If you do not see your python path listed, restart your R session and rerun the above commands in order.

3. Install SigProfilerAssignmentR using devtools:

>> library(devtools)
>> install_github("AlexandrovLab/SigProfilerAssignmentR")

4. If mutation calling files (MAF, VCF, or simple text files) are used as input, please load the package in the same R session and install your desired reference genome as follows:

>> library(SigProfilerAssignmentR)
>> install("GRCh37", rsync=FALSE, bash=TRUE)

This will install the human GRCh37 assembly as a reference genome.

SUPPORTED GENOMES

Other available reference genomes are GRCh38, mm9, mm10, and rn6. Information about supported genomes can be found at https://github.com/AlexandrovLab/SigProfilerMatrixGeneratorR

<a name="running"></a> Running

Assignment of known mutational signatures to individual samples is performed using the cosmic_fit function. Input samples are provided using the samples parameter in the form of mutation calling files (VCFs, MAFs, or simple text files), segmentation files or mutational matrices. COSMIC mutational signatures v3.3 are used as the default reference signatures, although previous COSMIC versions and custom signature databases are also supported using the cosmic_version and signature_database parameters. Results will be found in the folder specified in the output parameter.

>> library(SigProfilerAssignmentR)
>> cosmic_fit(samples, output, input_type='matrix', context_type="96",
              collapse_to_SBS96=TRUE, cosmic_version=3.3, exome=FALSE,
              genome_build="GRCh37", signature_database=NULL,
              exclude_signature_subgroups=NULL, export_probabilities=TRUE,
              export_probabilities_per_mutation=FALSE, make_plots=TRUE,
              sample_reconstruction_plots=FALSE, verbose=FALSE)

<a name="parameters"></a> Main Parameters

<a name="subgroups"></a> Signature Subgroups

When using COSMIC reference signatures, some subgroups of signatures can be removed to improve the refitting analysis. To use this feature, the exclude_signature_subgroups parameter should be added, following the sintax below:

exclude_signature_subgroups = ['MMR_deficiency_signatures',
                               'POL_deficiency_signatures',
                               'HR_deficiency_signatures' ,
                               'BER_deficiency_signatures',
                               'Chemotherapy_signatures',
                               'Immunosuppressants_signatures'
                               'Treatment_signatures'
                               'APOBEC_signatures',
                               'Tobacco_signatures',
                               'UV_signatures',
                               'AA_signatures',
                               'Colibactin_signatures',
                               'Artifact_signatures',
                               'Lymphoid_signatures']

The full list of signature subgroups is included in the following table:

<a name="examples"></a> Examples

Using mutation calling files (VCFs) as input

library(SigProfilerAssignmentR)

cosmic_fit(samples="test_data/vcf_input", 
           output="example_vcf",
           input_type="vcf",
           context_type="96",
           genome_build="GRCh37",
           cosmic_version=3.3)

Using a multi-sample segmentation file as input

library(SigProfilerAssignmentR)

cosmic_fit(samples="test_data/cnv_input/all.breast.ascat.summary.sample.tsv", 
           output="example_sf",
           input_type="seg:ASCAT_NGS",
           cosmic_version=3.3,
           collapse_to_SBS96=FALSE)

Using a mutational matrix as input

library(SigProfilerAssignmentR)

cosmic_fit(samples="test_data/txt_input/sample_matrix_SBS.txt", 
           output="example_mm",
           input_type="matrix",
           genome_build="GRCh37",
           cosmic_version=3.3)

<a name="copyright"></a> Copyright

This software and its documentation are copyright 2022 as a part of the SigProfiler project. The SigProfilerAssignmentR framework is free software and is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

<a name="contact"></a> Contact Information

Please address any queries or bug reports to Marcos Díaz-Gay at mdiazgay@health.ucsd.edu.