31 Detecting Similarities and Dissimilarities between Families of Bio-Sequences: A Neural Approach

In this paper we use a PCA neural network to detect similarities and dissimilarities between any two given families of biological sequences. Traditionally, PCA is a transformation technique used to reduce the dimensionality of a dataset, and transform it into a lower dimensionality space without any loss of information. In this paper we use PCA as a measure to detect similarities and dissimilarities between families of sequences. This is in contrast to detecting similarities between two sequences and between a sequence and a family. Of course, this method could be generalized for any datasets, and it will not be restricted for families of sequences. We propose a novel algorithm which can be used as similarity measure; we call it a PCA-neural network-based similarity measure. The performance of the proposed measure shows robustness and accuracy in similarity detection.