**We've moved!**

This site is now read-only. You can find our new documentation site and support forum for posting questions here.

Be sure to read our welcome blog!

# StrandOddsRatio interpretation

mkaram
LilleMember ✭

I would like to know how to interprete the StrandOddsRatio (SOR) obtained with HC in GATK.

Here is a summary of the SOR values that I obtained:

Min. 1st Qu. Median Mean 3rd Qu. Max.

0.001 0.598 0.725 0.819 0.895 7.643

Do the values around 1 indicate no or low strand bias? And the values close to 0 and greater than 1 indicate a strand bias?

Thank you.

## Answers

Hi there, have you looked at the SOR annotation documentation? See https://www.broadinstitute.org/gatk/guide/tooldocs/org_broadinstitute_gatk_tools_walkers_annotator_StrandOddsRatio.php

Dear Geraldine, to be clear in my question, do we get log(R+1/R) in the SOR column of the vcf file?

@mkaram

Hi,

Yes, it is the ln(R+1/R). I am going to document this properly in the near future.

-Sheila

Dear Sheila, thank you for your answer. I still don't understand something: the (R+1/R) function has a minimum of 2, meaning that ln(R+1/R) should be greater than ln(2) = 0.693. How can I explain my SOR values that are between 0.001 and 0.693? Thank you.

@mkaram

Hi,

The final calculation is not just ln(R/1+R). It is ln(ratio*(R/1+R)). The ratio calculation is:

((ref forward / ref reverse) / (alt reverse / alt forward)) + ((ref reverse / ref forward) / (alt forward / alt reverse))

You add 1 to each of the counts in order to avoid dividing by 0 and to account for any potential bias.

I hope this helps.

-Sheila