in reply to Can't think of an algorithm to go..

if for the letter B of SEQ, the first LBL line says I, the second I, the third M and the fourth I, we will use I in the consensus label line.

I don't see a letter B in the SEQ line of your example, but I guess you mean the following (again using your example):

It is not clear how to associate, say, the letter K, because the first K has only I in the LBL lines, while a later one has only M. also, what is the consensus if there is the same number of, say, I and O? Maybe you could specify your task in a more precise way?

-- 
Ronald Fischer <ynnor@mm.st>