Seeking Significant Oligomers via Set Partitions Expected Count

In order to determine significance of word counts of DNA sequences, it is of first importance to develop a baseline comparison so that the non-randomness of the observed word count can be measured. We developed a novel measure of oligomer expected count using the concept of set partitions. This expected count provides a baseline reference to reveal non-random DNA sequences. Non-randomness of oligomers is evaluated in terms of the amount of deviation from the derived expected count. As a consequence, the ratio of the observed count to the expected count will indicate the degree of underor over-representation of the oligomers. The usefulness of the method is demonstrated when applied to two human chromosomes and an artificially generated random chromosome. Underand over-represented oligomers are revealed in the human chromosomes but not in the random chromosome.