Bash edit numbers in a column based on the value beneath it

Question

I have a file with two columns

A *
B 2
C 4
D *
E *
F 6
G 4
H 5

The asterisk means that it is in the same cluster with the row beneath it with