Search
Question: mutational signatures problems
0
2.5 years ago by
jolligoodfellow0 wrote:

Dear Forum,

I am working on a somatic signatures project and I would like to make somatic signatures of the 80 Uveal Melanoma samples on the TCGA site but I encounter some problems.

First: how can I transform .MAF files from the TCGA site in to .VCF file in order to use the vr_A = readVcfAsVRanges(vcf_A_path, "GenomeA") command in Rstudio?

Second: how can I use and upload the reference genome in my Rstudio used for the Uveal Melanoma. I know they used the HG19 build available on this link https://browser.cghub.ucsc.edu/help/assemblies/

Third: after having this worked out I would like to follow the steps described in the somatic signature vignette eventually leading to the use of this command: plotMutationSpectrum(sca_motifs, "study")

Probably I am tackling this problem the other way around but because Uveal Melanoma is not available in the online data after entering this code from the vignette sca_metadata = scaMetadata() I have to try something else.

I am a beginner in this field and I hope you are willing to answer (some) the questions and help me out.

modified 2.5 years ago by Julian Gehring1.3k • written 2.5 years ago by jolligoodfellow0
0
2.5 years ago by
Julian Gehring1.3k
Julian Gehring1.3k wrote:

First: how can I transform .MAF files from the TCGA site in to .VCF file in order to use the vr_A = readVcfAsVRanges(vcf_A_path, "GenomeA") command in Rstudio?

My personal advice would be to avoid MAF files and try to get access to VCF files whenever possible, since MAF files are not very well defined with regard to their format. TCGA may have tools for such a convertion, I'm not aware of a standard solution in R/Bioc. The alternative would be to do it yourself, you can have a look at the functions in the SomaticCancerAlterations package to get an idea - but please understand I don't expect these import functions to work on MAF files in general and are not supported.

Second: how can I use and upload the reference genome in my Rstudio used for the Uveal Melanoma. I know they used the HG19 build available on this link https://browser.cghub.ucsc.edu/help/assemblies/

Bioconductor has hg19 packaged as one of their BSgenome packages, so you can install it this way.

Third: after having this worked out I would like to follow the steps described in the somatic signature vignette eventually leading to the use of this command: plotMutationSpectrum(sca_motifs, "study"). Probably I am tackling this problem the other way around but because Uveal Melanoma is not available in the online data after entering this code from the vignette sca_metadata = scaMetadata() I have to try something else.

The data from the vignette comes from the SomaticCancerAlterations package, which all the analysis that you are interested in is performed by the SomaticSignatures package. The first steps only concern get/preparing the data for the analysis - if you use your own data, you may have to take different steps here. For example, scaMetadata() contains metadata for the 8 TCGA studies from the SomaticCancerAlterations package, but not those of additional studies.