speedup when parsing cigarstring in alignment

gistfile1.txt

    @property
    def alignment(self):
        r_index = 0
        q_index = 0
        r_seq = self.reference[self.reference_begin: self.reference_end + 1]
        q_seq = self.query[self.query_begin: self.query_end + 1]
        r_line = m_line = q_line = ''
        for (op_len, op_char) in self.iter_cigar:
            op_len = int(op_len)
            if op_char.upper() == 'M':
                ref_piece = r_seq[r_index: r_index + op_len]
                query_peace = q_seq[q_index: q_index + op_len]
                r_line += ref_piece
                q_line += query_peace
                m_list = [] # faster with "".join([list of str]) instead of +=
                for (r_base, q_base) in zip(ref_piece, query_peace):
                    # XXX: ambiguity codes? matrix match?
                    
                    if r_base.upper() == q_base.upper():
                        m_list.append('|')
                    else:
                        m_list.append('*')
                match_seq = "".join([char for char in m_list])
                m_line += match_seq
                r_index += op_len
                q_index += op_len

            elif op_char.upper() == 'I':
                # insertion into reference
                r_line += '-' * op_len
                m_line += ' ' * op_len
                q_line += q_seq[q_index: q_index + op_len]
                #  only query index change
                q_index += op_len

            elif op_char.upper() == 'D':
                # deletion from reference
                r_line += r_seq[r_index: r_index + op_len]
                m_line += ' ' * op_len
                q_line += '-' * op_len
                #  only ref index change
                r_index += op_len
        return (r_line, m_line, q_line)