Does the bias have a direction?

Nc plot can indicate if there's a bias, but not which way 
it goes (eg. whether codon usage is more similar to host).

We need another kind of analysis. 

Correlation of individual codon usage tables between two 
species: simplistic and not very powerful, but it can 
show us direction of codon bias.

Obtain correlation coefficients based on frequency /1000
of individual codons for two species, get r value.

no correlation   0  <-  r  ->  1  perfect correspondance