Is it possible to use telomere to telomere instead of hg38?

[JanSliacky]

We are started to use mapping t2t instead of hg38. I see there is the library hg38 necessary for running the functions. We need the rds file for respective human genome. I saw the source in the official sigminer pages with genomes up to hg38. Would it be possible to use t2t?

[ShixiangWang]

Hi @JanSliacky , it's a good point. Could you help check if https://bioconductor.org/packages/release/data/annotation/html/BSgenome.Hsapiens.NCBI.T2T.CHM13v2.0.html could fit your needs? If so, I could add some code to support the genome.

[ShixiangWang]

Ok. These are just index files. I will do some search work in my spare time and see if it is possible.

In the current stage, I recommended you generate the MAF file by yourself (e.g., vcf2maf, also Maf data is always more recommended in sigminer, if you use a vcf as input, sigminer generates a fake Maf format data), and then read the data with read_maf and generate the matrix for signature extraction with sig_tally.

[JanSliacky]

Thanks for hint. I tried to use vcf2maf. However, it is a real pain to use it. I used dockerized version: https://hub.docker.com/r/opengenomics/vcf2maf but the perl script demands quite a lot files for reference. Some of them I saw for the first time.

Do you have any documentation or best pracise for using vcf2maf?

[ShixiangWang]

Hi, indeed the vcf2maf is a huge based on vep. You may want to try https://annovar.openbioinformatics.org/en/latest/, If you’re using ANNOVAR for variant annotations, maftools has a handy function annovarToMaf for converting tabular annovar outputs to MAF.

https://github.com/IARCbioinfo/table_annovar-nf here is a pipeline, I don't knonw if it fits the T2T. You can also see https://github.com/WGLab/Workshop_Annotation#install-annovar

I don't have much experience on the vcf to maf.

[ShixiangWang]

Hi @JanSliacky

I just pushed code updates to support T2T, can you try installing the latest version and test it?

The code you used should work. If there are any bugs, please report them to me.

Do you think this code would work?

mtf <- data.frame()
for (i in 1:length(vcfs)) {
  m <- read_vcf(vcfs[i], keep_only_pass = FALSE, genome_build = c("T2T")
  mt_tally <- sig_tally(
    m,
    ref_genome = "BSgenome.Hsapiens.NCBI.T2T.CHM13v2.0",
    genome_build = "T2T",
    use_syn = TRUE,
    cores = 16
  )
  mtf <- rbind(mtf, mt_tally$nmf_matrix)
}