Automated chromosome analysis is now becoming routine in most human cytogenetics laboratories. It involves both processing and analysis of digital images and has been developed because of the demandby cytogeneticists. Over the years, many techniques have been introduced for the automatic segmentation and classification of chromosome images, of which only a few are included in the available commercial systems. Today, advances in chromosome imaging techniques, especially in multispectral imaging, lead the way for the development of new and improved methods for the location, segmentation and classification of chromosome images by exploiting the color information. In this chapter the authors describe methods which have been already developed for automated chromosome analysis.
Chromosomes are structures that contain genes, which store in strings of DNA all the data necessary for an organism’s development and maintenance. They contain vast amounts of information; in fact each cell in a normal human being contains 46 chromosomes which have bits of information (Thompson, 1992). Chromosomes can only be examined visually during cell division (mitosis). They are extremely long and thin which make them essentially invisible. However, during the metaphase stage of mitosis, they contract and become much shorter (around 2–10μm) and wider (around 1–2 μm diameter), (Figure 1(a)). At this stage, they can be stained to become visible and can be imaged by a microscope.
(a) A slide of grayscale banded chromosomes and (b) their karyotype.
Chromosome analysis is the procedure from which chromosomes are photographed during cell division and then are assigned to each class. This procedure is called karyotyping, where chromosomes are aligned in pairs in a tabular array as it is shown in Figure 1(b). Karyotyping is a useful tool to detect deviations from normal cell structure. Examples include peripheral blood, bone marrow, amniotic fluid, and products of conception. Normal human somatic cells have 46 chromosomes: 22 pairs of autosomes (chromosomes 1-22) and two sex chromosomes. Females carry two X chromosomes (46, XX), while males have a X and a Y (46, XY). Germ cells (egg and sperm) have 23 chromosomes: one copy of each autosome plus a single sex chromosome. This is referred to as the haploid number. One chromosome from each autosomal pair plus one sex chromosome is inherited from each parent. Mothers can contribute only an X chromosome to their children, while fathers can contribute either an X or a Y. Abnormal cells may have an excess or a deficit of chromosomes and/or structural defects which depict an exchange of genetic material.
Chromosome abnormalities can be very complex. There are two basic types of abnormalities: numerical and structural and both types can occur simultaneously. The most obvious abnormality is an unusual number of chromosomes. Having only one type of chromosome is a monosomy, such as Turner’s syndrome, in which there is only one X chromosome and no Y. Having three chromosomes is a trisomy, such as Down’s syndrome, in which there are three Type-21 chromosomes.
There can also be duplications of genetic material within a chromosome and translocations where two chromosomes exchange genetic information. The Philadelphia chromosome results from a translocation in the 9th and 22nd chromosomes. This is often associated with chronic myelogenous leukemia (Nowell, 1960). Detecting these abnormalities is vital because they are reliable indicators of genetic disease and damage. Chromosome abnormalities are particularly useful in cancer diagnosis and the related research (Gray, 1992).
Digital imaging has contributed to cytogenetics instrumentation reducing the workload in clinical labs and producing quantitative data for both research and diagnosis. The last few decades we have seen continuous endeavors in (a) the development of innovative image acquisition and enhancement methods on technologies that exploit our knowledge of the molecular basis of cancer or other diseases, and (b) the integration of these emerging genomic technologies with traditional imaging methods for more effective solutions for health care delivery. In this chapter we introduce the reader to the state of the art for automated methods in chromosome analysis.
The methods presented below are divided into two main categories based on the type of the image which is used.
Key Terms in this Chapter
Chromosome: A chromosome is a continuous piece of DNA, which contains many genes, regulatory elements and other nucleotide sequences.
Centromere Index: The centromere index is defined as the ratio of the length of the short arm of the chromosome divided by the length of the other arm.
Machine Learning: As a broad subfield of artificial intelligence, machine learning is concerned with the design and development of algorithms and techniques that allow computers to “learn”.
Centromere: The centromere is a region, found in the middle of the chromosome, involved in cell division and the control of gene expression.
Class ification: The process of deriving a mathematical function that can predict the membership of a class based on input data.
Watershed: The segmentation based on watershed designs is a family of segmentation methods that consider an image as a topographic relief the flooding of which is simulated.
Complete Chapter List
Themis P. Exarchos, Athanasios Papadopoulos, Dimitrios I. Fotiadis
Ioannis Dimou, Michalis Zervakis, David Lowe, Manolis Tsiknakis
William Hsu, Alex A.T. Bui, Ricky K. Taira, Hooshang Kangarloo
Spyretta Golemati, John Stoitsis, Konstantina S. Nikita
Christos V. Bourantas, Katerina Naka, Dimitrios Fotiadis, Lampros Michalis
Stavroula Mougiakakou, Ioannis Valavanis, Alexandra Nikita, Konstantina S. Nikita
Marotesa Voultsidou, J. Michael Herrmann
Dimitrios C. Karampinos, Robert Dawe, Konstantinos Arfanakis, John G. Georgiadis
Dimitrios G. Tsalikakis, Petros S. Karvelis, Dimitrios I. Fotiadis
Katia Marina Passera, Luca Tommaso Mainardi
Lena Costaridou, Spyros Skiadopoulos, Anna Karahaliou, Nikolaos Arikidis, George Panayiotakis
E. Kyriacou, C.I. Christodoulou, C. Loizou, M.S. Pattichis, C.S. Pattichis, S. Kakkos
Marios Neofytou, Constantinos Pattichis, Vasilios Tanos, Marios Pattichis, Eftyvoulos Kyriacou
Thomas V. Kilindris, Kiki Theodorou
Ioannis Tsougos, George Loudos, Panagiotis Georgoulias, Konstantina S. Nikita, Kiki Theodorou
Evanthia E. Tripoliti, Dimitrios I. Fotiadis, Konstantia Veliou
Anastasios Koutlas, Dimitrios I. Fotiadis
Aristotelis Chatziioannou, Panagiotis Moulos
Nikolaos Giannakeas, Dimitrios I. Fotiadis
Petros S. Karvelis, Dimitrios I. Fotiadis
O. Lezoray, G. Lebrun, C. Meurie, C. Charrier, A. Elmotataz, M. Lecluse
Michael Haefner, Alfred Gangl, Michael Liedlgruber, A. Uhl, Andreas Vecsei, Friedrich Wrba
C. Papaodysseus, P. Rousopoulos, D. Arabadjis, M. Panagopoulos, P. Loumou
Alexia Giannoula, Richard S.C. Cobbold
Valentina Russo, Roberto Setola
George K. Matsopoulos
Alberto Taboada-Crispi, Hichem Sahli, Denis Hernandez-Pacheco, Alexander Falcon-Ruiz
C. Delgorge-Rosenberger, C. Rosenberger
Charalampos Doukas, Ilias Maglogiannis