Data interpretation

[Anahit19]

Dear Derek,

I am using the BIGDAWG package to run haplotype analysis for my HLA data.

As a result, for the alleles of C and B loci I have no significant association.

The only allele combination which differs is B3502 and C0401, which are supposed to be in linkage.

Could you please help with the output interpretation. What does the Inf OR mean?

Also is it possible to conduct Ordinal regression analysis for three disease subgroups with your package?

Thank you very much,

Regards, Anahit

[IgDAWG]

Hi Anahit,

An OR value of inf (meaning "infinity") for a given variant, which I think in your case is the B*35:02~C*04:01 haplotype, means that there are no observations of that variant in the controls, so the OR value essentially too high for BIGDAWG to calculate (essentially infinite); think of it as akin to a divide-by-zero result on a calculator. Note that the upper bound of the 95% confidence interval will also be inf.

When there are no observations of a variant in the cases (and there are sufficient observations in controls to avoid binning) the OR value is 0.

To illustrate this, I modified the HLA_data data frame bundled with BIGDAWG and set the number of A33:01:01 alleles to 0 in controls, and the number of DRB1*11:03 alleles to 0 in cases.

Here is a marked-up and edited View() of the OR table for those results:

And here is a marked-up and edited View() of the table of counts:

I hope this helps interpret your BIGDAWG results.

-- Steve Mack

[IgDAWG]

Oh. I should add that we don't currently have a regression function for BIGDAWG, but that will be coming in a release in the near future.

[Anahit19]

Thank you very much!