A Combination of Kernel Methods and Genetic Programming for Gene Expression Pattern Classification

摘要

The rapidly emerging field of quantitative proteomics has established itself as a credible approach for understanding of the biology of whole organisms. Classification of proteins according to the level of their expression during a particular process allows discovering causal relationships among genes and proteins involved in the process. In this paper, we would like to propose a new algorithm for pattern classification, allowing for extraction of user defined patterns from a database of kinetic gene expression profiles. This algorithm is a combination of kernel methods and genetic programming. The algorithm was tested on publicly available transcriptomic and proteomic time series datasets and the results showed that the algorithm could find all similar patterns in the database with very low misclassification rate.