Hi Pascal,
In 3DNA and DSSR, the classification of Z-DNA is on double-stranded helices, not single strands. Furthermore, in 3DNA (analyze), the classification is on the whole structure, whilst in DSSR, it is on each dinucleotide step.
Regarding the Z-DNA CpG motifs, I am glad that you paid attention to the O4'(C)...G capping interactions. From early on, I realized that it is not just the OP1/OP2 atoms of the phosphate groups that can make capping interactions, and that's why I implemented them all in DSSR. As documented in the manual, you need to specify the --more option to see the capping interactions.
HTH,
Xiang-Jun