Думаю гаплотипы AS убирать не стоит, хотя это Ваше дело.
Мне интересно взглянуть на гаплотипы I1 у эрзи, мокши и чувашей, хотя мне кажется они будут схожи с гаплотипами русских Пензы и Н.Новгорода.
Removing AS makes the figure much clearer. A pre-analysis with modals including also AS-types is a good idea and useful to do before splitting the picture.
I am just now learning how to define networks. I think that one most relevant issue is how key markers identify different populations, how good modals we have in this meaning. The key markers found by the modal analysis should get more priority.