Determine strand info of somatic mutations

Question

[whb]

Hi all,

How do I get the strand (transcribed/non-transcribed) information of somatic mutations? I have called the mutations with Mutect2 and annotated them with SnpEff and SnpSift but I cannot figure out how to find which strand a mutation is coming from in the vcf file.

Thanks

[cpad0112]

variants (records) on vcf are always denoted on + strand (same strand as reference genome). Unless there is a strand information in vcf, it is difficult to know the stranded information from VCF for each mutation.

Answer 1

[ATpoint]

DNA is double-stranded. Hence mutations are double-stranded and not strand-specific.

[whb]

Thanks for the reply. I am wondering if I have missed any info because a few papers compare the number of substitution mutations on the transcribe and non-transcribed strand e.g. Figure 1D in https://pubmed.ncbi.nlm.nih.gov/25999502/ or did I misunderstand something? Thanks again.

[ATpoint]

They seem to look at the exact nucleotide change. So the position in the chromosome is the same (as it is double-stranded) but if a gene is on top-strand then the nucleotide would e.g. be A and if it was on the bottom it would be T. You could intersect your mutations with a GTF file, retain those that overlap a gene, and then filter for the strand information. Next, use something like bedtools getfasta to extract the exact nucleotide for every mutation, based on the top strand, and if the intersection from the step before indicated minus strand then convert the base to its complement. Does that make sense for you? What is your coding level, can you write something like this or do you need a end-to-end tool?

[whb]

Hi ATpoint thank you for the reply. I am a beginner in coding so it would be great if you could suggest some end-to-end tools. From what you have suggested, should I start with intersecting vcf files with GTF? or bam/fastq? and I am not sure how to extract strand inforamtion?