It seems that OP's problem is an easy special case of the LCS problem. We have one big string S and 2 substrings (seq1, seq2) with known coordinates within S:

S:    CTGCCACCGCTGTGTTTCGGCCGGCGAXXXXXXXXXXXXXXXXXXXXX
seq1: CTGCCACCGCTGT (Pos: 0)
seq2:      ACCGCTGTGTTTCGGCCGGCGA (Pos: 5)
[download]

So it's a trivial O(1) algorithm to calculate the overlap with the substring positions and lengths.