Cancerous/Disease DNA Prediction Using Fixed Length Motifs/Frequent Patterns Matching

In the radical field of bioinformatics, one very interesting and rather concerning area of research is predicting cancer infected gene from a set of samples of species DNA. This field is quite a challenging one considering the limited knowledge on how cancers affect gene of species and the pattern of mutation are not always the same. Gene prediction can be effectively done through several techniques like frequent pattern mining, neural networks or sequence alignment. These traditional approaches were able to predict to a very small limit. In this paper a new method using frequent patterns/motifs is shown that can be a new strategy for prediction of gene in a DNA. As the motifs in a DNA are the conserved region, so it's more appropriate to be used for gene predication and alignment. The new method proposed in this paper includes the sampling of fixed length motifs from a sequence of reference genome and finally other samples are aligned against the more frequent motifs to establish their relevancy to the reference genome.