Discrepancy between mutalilsk result and FIREVAT result

[Yoon-Seob-Kim]

Hello

Using TCGA data vcf file, I found that there is a discrepancy regarding number of point mutation (in original VCFs) using FIREVAT result from RunFIREVAT function and mutalisk result from web page.

When i check number of point mutation manually, i found that mutalisk result was in concordance with manually counted number of point mutation , but FIREVAT result had fewer point mutation and several point mutations were classified as "Unknown".

The genome version has been inputted correctly as hg38, and i used FIREVAT with the vesion of 0.5.9.

Could you check there is any difference between two algorithms?

Thank you.

[andyjslee]

Hello @kysbbubbu,

Thank you for using FIREVAT.

In short, the difference in the results you obtained from FIREVAT and Mutalisk is as expected. This is because FIREVAT initially filters out point mutations that do not have the desired filter parameter values. You will find that these mutations end up in the "*unknown.vcf" file. Please note that if at least one of the filter parameter values is missing for a given point mutation (based on the configuration file supplied), the same mutation will be classified as "unknown". On the other hand, Mutalisk considers all point mutations listed in your VCF file.

I hope this answers your question. Please feel free to let us know if you have any additional questions.

-Andy