摘要
Weight matrix models for signal sequence motif are simple. A main limitation of the models is the assumption of independence between positions. Signal enhancement is achieved by taking the total likelihood as the objective function for maximization to cluster sequences into groups with different patterns. As an example, the initial and terminal signals for translation in rice genome are examined.
基金
the Special Funds for Major National Basic Research Projects，国家自然科学基金，Research Project 248 of Beijing
