Using these independent variables for each variant of AKAP11 in
SCHEMA and BipEx, we take the frequencies of the amino acids coded for
by each of the variants with the highest 10% probabilities of being
placed in j = 4 (i.e., most likely to cause diagnosis of SCZ and
BPD). If there were no correlation with AKAP11 , we would assume
a null hypothesis wherein there is no difference between the amino acid
distribution of the highest risk AKAP11 variants and the natural
distribution of amino acids in humans, demonstrating the inability of
the gene to be damaging to non–random amino acids. A chi–square
goodness of fit test (χ2) is used to determine if
there is any statistical deviation from the amino acid frequencies
observed against natural amino acid frequencies, given that the data
adhere to the standard assumptions of χ2.
Particularly, variables are ordinal, mutually exclusive, and absolute
\citep{McHugh2013}.