How to choose variants for viewing in IGV?

Question

[Martina]

Hello guys, do you have some "verified" strategy about how to choose of variants for viewing in IGV? :)

Thank you very much, Martina

[ATpoint]

Hi Martina, please put some efort into the question by providing the necessary details. What kind of variants are you referring to (short, large, SNP, Indel, structural) and what exactly do you want to do with it. What is the input format and most importantly what is the criterium to choose variants?

[Martina]

Sure, I am processing exome tumor-normal paired data and I am searching for somatic variants. We are not focusing on structural variants because the DNA is amplified. I have bam files created by BWA-mem and vcf files from freebayes filtered out with several filters. Now I would like to make an overview of what we have and I was also asked to choose several variants to take a look at in IGV, so I would like to cover some reasonable spectra of it. But how should I choose that spectra? Probably there should be some amount of indels, SNVs, variable coverage, variable position on genome.. Something else? What is a reasonable amount of them?

Answer 1

[dganiewich]

Dear Martina, I believe the answer is as broad as the question...

I guess you obtained a long VCF and don't know where to start. As a general approach, you can start from most probably associated with disease to less probable by prioritizing variants. You could look at the type of tumor you are working with and based on the phenotype observed go and look firstly at the genes relevant for those, which of course relies on your expertise on the disease (and bibliography). If nothing is found, then search for less associated but still known genes and so on.

I also work with paired tumor-normal exome and after filtering based on the population frequency, effect and genes relevant for cancer we manually review all of the variants on IGV. In my case, for pediatric cancer St. Jude's Pecan Pie (https://pecan.stjude.cloud/) has been of great help because it prioritizes and annotates the variants, and also Cancer Genome Interpreter (https://www.cancergenomeinterpreter.org/) prioritizes in tiers, take a look at their documentation; but many other annotators will give you great useful extra information to consider.

Don't know if this answers your question, but the thins is that there is no systematic way of reviewing variants that works for every sample and every disease, otherwise someone would have already coded it :)

Hope this helps,

Best,

Daiana

[Martina]

Dear Daiana, that's exactly that kind of answer I needed, thank you very much!! I also found this great tutorial about what sequencing artefacts look like in IGV, https://bioinformatics-core-shared-training.github.io/intro-to-IGV/InspectingVariantsInIGV.html and this interesting paper https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5678989/ so also other people who will find this question useful can have it on site. Also, if somebody knows about some useful IGV viewing studying material, don't be shy to post here ;)

Have a nice day,

Martina