Transition/Transversion Ratio From Fastq Files

Question

[venks]

Hi,

I was doing QC for my paired-end exome data which is in FASTQ format. I used FASTQC for the QC but I am more interested in looking at the Transition/Transversion ratio. Is there any tool that can give us the transition/transition ratio?

Thanks

[Manu Prestat]

Don't you need to compare similar sequences (then assumed as homologous) to derive substitution type estimates?

[Gabriel R.]

I am very confused, fastq gives you raw sequences. to get transition/transversions, you need aligned reads to a reference.

Answer 1

[Chris Miller]

You're several steps short of looking at Ti/Tv ratios. I'm going to talk about this in the context of cancer, because it's what I'm most familiar with. Your steps will be something like:

• use an aligner (like BWA) to align both your tumor reads and normal reads to the reference genome.
• Use a variant caller (samtools, Varscan, strelka, etc) to identify somatic snvs
• determine whether each of these sites was a transition or transversion, then calculate the rates.

[venks]

Thanks Chris. So which tool you think can be used to calculate the ti/tv rate?

[Chris Miller]

Once you get your list of mutations, it will contain the reference and variant bases for each. With that information it's very straightforward to write a script that counts whether each is a transition or transversion. http://www.mun.ca/biology/scarr/Transitions_vs_Transversions.html

[venks]

Perfect. Thanks :)

[William]

Once you have a vcf produced by a variant caller you can run GATK VariantEval: http://www.broadinstitute.org/gatk/gatkdocs/org_broadinstitute_sting_gatk_walkers_varianteval_VariantEval.html

It will produce statistics on your variant, including Ti/Tv ratio.

Answer 2

[zam.iqbal.genome]

Also, you may know this, but worth saying, Ts/Tv varies between species (not sure what species you are using).

[venks]

I am using Human sequence data

Answer 3

[rob234king]

If you have vcf file to annotate use snpEff and in the html report it will show amongst other things the ratio you want.