[igraph] Random graph based on probabilities derived from sample graph

[capitano . nemo]

Hi,
I have mapped directed relationship among a sample of people and I would 
like to create a much larger graph with probability of having a 
relationship between every pair of people derived from my sample data. 
Also my probability model should take into account that the probability 
of event Alice -> Bob is not totally independent from the probability of 
event Alice <- Bob.
I guess then that I should be able to define a probability model based 
on four type of events for each pair of people:
1) p of "No relationship";
2) p of "A -> B";
3) p of "A <- B";
4) p of "A <-> B".
Is that right? How should I compute the model from my graph? Is there 
any function I should use?

I am moving the first steps on this, so every suggestion is more than 
welcomed.
Thanks
Francesco