Automatic chromosome classification and chromosome abnormalities identification based on dynamic time warping

Abstract

Identification of chromosome abnormalities is an essential part of diagnosis and treatment of genetic disorders. However, manual chromosome analysis is a slow and laborious operation that requires qualified personnel. Nowadays, there is a lack of automatic software able to deal with abnormal chromosome images. In this study, the dynamic time warping algorithm has been adapted to the identification of chromosome banding patterns in order to overcome the problems due to the non-rigid nature of chromosomes. Firstly, a pattern classifier has been developed to perform the classification of normal human chromosomes. This classification method has the main advantage to require only a few training samples in comparison with the traditional chromosome classifiers. Secondly, a procedure is proposed to automate the study of abnormal chromosomes by using G-banding and spectral karyotyping images. This method has been successfully applied to the mapping of chromosome translocations, duplications and deletions in solid tumor cells.