Characterization of chromatin distribution in cell nuclei

Abstract

In this paper we develop four measures to describe the distribution of nuclear chromatin. These measures attempt to describe in an objective and meaningful way the heterogeneity, granularity, condensation, and margination of chromatin in cell nuclei. Starting with a high-resolution digitized image of a cell where the nuclear pixels have been identified, the four measures may be rapidly estimated. The range of each is derived and the interpretation of the measures in the context of chromatin compaction and distribution is developed. Implementation issues such as sampling density, thresholding and subsequent pre-processing, and algorithmic complexity are discussed.Key terms: Quantitative microscopy, image processing, texture measures, pattern recognition, image measurementIn this paper we present four measures that we have developed to quantify and characterize the distribution of chromatin in the nuclei of stained cells. These measures are based upon a simple model for the way chromatin compacts in a cell nucleus and how that compaction is reflected in verbal descriptions of nuclear "texture." Use of the word texture immediately sug gests a range of possibilities for quantification and it is not our purpose here to go into a lengthy review of the texture literature. What has guided us in this study, however, has been a search for ways of describing chromatin distribution that clearly relate to the underlying processes that cause change in the appearance of cell nuclei. Thus we start from a somewhat different perspective then the mathematical texture parameters of Haralick (5), Galloway (3, or Pressman (16). METHODS ModelingWe begin with the idea that a cell nucleus (as pictured in Figs. 1A,B) has a constant amount of DNA except when the cell is synthesizing DNA or about to divide.Thus we are considering cells with a normal, diploid (212) DNA content-GO and G1 cells. In a typical cell population such cells will account for at least 85% of all randomly sampled cells (14). It is precisely these cells that we are interested in characterising with respect to chromatin distribution; cells with a DNA complement above the 2c level are easily found by measuring their DNA content. If a difference exists between the nuclear chromatin pattern in two similar cells, then, according to our model of constant DNA, it must be due to a redistribution of the chromatin within the nucleus. This is illustrated in Figure 2A-D where all four cell images have precisely the same total nuclear optical density (proportional to DNA) but the distributions have been artificially altered.We describe the possible chromatin variations in terms of three linguistic attributes: heterogeneity, granularity, and margination. The first, heterogeneity, refers to whether the nuclear chromatin is homogeneously distributed throughout the nucleus or condensed into granules.If the chromatin is condensed, then the description of granularity assumes a useful role. While Figures 2C and D both exhibit an artificial granularity, the size of the "granules" differs s...

Karyometry of the Colonic Mucosa

“…The number of runs of different lengths and in different absorbance ranges provides a set of features. Finally, there are features descriptive of the local arrangement of multiple pixels of given gray values: chromatin condensation (referred to as pixel absorbance condensation); chromatin or pixel absorbance clumpness; and pixel absorbance heterogeneity and homogeneity (25). The full set of features captures a broad range of information that maximizes the opportunity to detect nuclear changes of significance in cancer progression.…”

Karyometry detects subvisual differences in chromatin organization state between cribriform and flat high-grade prostatic intraepithelial neoplasia

“…These are related to nuclear area, total optical density and chromatin distribution. [11][12][13] Table 1 gives a sample list of features used in the discriminant function analyses and in the unsupervised learning program P-index [11][12][13] (see below) (all features are given in relative units of measure) (the values in parenthesis refer to an arbitrary code number with which the feature is identified in the computer program). The P-index groupings are based on two composite features: the discriminant function I score, and the nuclear abnormality.…”

“…[11][12][13] These values express the relative deviation of each feature from 'normal', as assessed from the set of 'normal' reference nuclei. These features have the advantage that they are based on a relative, 'internal' standard.…”

A Technique for Extraction of Diagnostic Data from Cytological Specimens

“…The second approach, textural, is based on the statistical characteristics of chromatin arrangement and related to analysis of the regularities of chromatin structure. Applied in practice methods for textural analysis use grey level dependency matrices [8], co-occurrence, run-length features, rice-field operators, and watersheds (topological methods) [10], heterogeneity, clumpiness, margination, and radius of particles [12] (the Mayall/Young features), invariant features (polinomial invariants).…”