![](bg1.png)
AUTOMATED DETECTION OF INTERPHASE AND METAPHASE
NUCLEI IN THE FISH IMAGES
Jan Schier
1
, Bohumil Kov
´
ar
1,2
and Eduard Koc
´
arek
3
1
Institute of Information Theory and Automation of the ASCR
Pod vod
´
arenskou ve
ˇ
z
´
ı 4, Prague, Czech Republic
2
Faculty of Transportation Sciences, Department of Applied Mathematics, Czech Technical University in Prague
Na Florenci 25, Prague, Czech Republic
3
Department of Biology and Medical Genetics, 2nd Faculty of Medicine, Charles University Prague
V
´
Uvalu 84, Prague, Czech Republic
Keywords:
FISH analysis, Fluorescence microscopy, Image analysis, Image segmentation.
Abstract:
The fluorescence in-situ hybridization (FISH) belongs to the most common cytogenetic methods and is widely
applied in routine clinical genetic diagnostics. We are paying attention to FISH analysis of chromosomal ane-
uploidies – the deviations from chromosomal number. Such analysis is based on evaluation of up to several
hundreds of microscopic images. Computer support for this process includes using methods of image process-
ing and data mining. In this paper, we focus on the image processing part in more detail: first, the properties
of FISH images are reviewed, then, the processing flow is outlined. Our aim is to find the interphase and
metaphase nuclei and the hybridization signals contained in the image. A simple method using the raw and
central moments of detected objects as measures to distinguish between the two types of nuclei is proposed.
1 INTRODUCTION
Fluorescence in situ Hybridization (FISH) is a tech-
nique used to visualize the location of specific DNA
sequences within the nucleus. It belongs to the most
common of molecular cytogenetic methods and is
widely applied in routine clinical genetic diagnostics.
The FISH technique concerns application of a flu-
orescently labeled DNA or RNA probe onto a biolog-
ical sample containing the target DNA or RNA from
the patient (e.g. as tissue sections, separated cells,
isolated interphase nuclei or chromosomal metaphase
spreads). In the process of hybridization, the probe
attaches to its complementary sequences on the target
DNA or RNA molecule. Under the fluorescence mi-
croscope we observe the sequence with the attached
probe as a strikingly luminous hybridization signal (a
fluorescent dot or an area within the biological struc-
ture). Details of the principles of the fluorescence mi-
croscopy can be found e.g. in (Rittscher et al., 2008)
FISH is also a useful diagnostic tool in repro-
ductive medicine, oncology, preventive medicine, and
other medical branches – examples can be found in
(Beatty et al., 2002) or (Varella-Garcia, 2003).
We are paying attention to FISH analysis of chro-
mosomal aneuploidies – i.e., deviations from chromo-
somal number. These abnormalities are responsible
for well-known clinically significant syndromes such
as Down’s syndrome (trisomy 21) or Turner’s syn-
drome (monosomy X), Klinefeter’s syndrome (XXY
constitution) – see (Harper, 2004), (Roizen and Pat-
terson, 2003), (Anner
´
en, 2007), and (Visootsak and
Graham, 2006).
To be able to conclude examination of a single pa-
tient, a set of tens or hundreds of images have to be
evaluated. An automated data acquisition and pro-
cessing is needed to achieve acceptable accuracy and
reproducibility of results; manual processing would
be time consuming and subjective.
Our paper is focused on the task of image segmen-
tation, that is, on finding and evaluation of the objects
of interest — the interphase cell nuclei, metaphase
spreads and the hybridization signals. We first re-
view the properties of images resulting from the FISH
analysis in Section 2 and propose an image process-
ing system in Section 3. In Section 4, we discuss the
edge-based segmentation technique used for identi-
fication of cell nuclei and detection of hybridization
signals. In Section 5, we describe the experiments
and, finally, Section 6 concludes the paper.
347
Schier J., Kovár B. and Kocárek E..
AUTOMATED DETECTION OF INTERPHASE AND METAPHASE NUCLEI IN THE FISH IMAGES.
DOI: 10.5220/0003797703470350
In Proceedings of the International Conference on Bioinformatics Models, Methods and Algorithms (BIOINFORMATICS-2012), pages 347-350
ISBN: 978-989-8425-90-4
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)