Granulometric analysis of spots in DNA microarray images.

Area of Science:

• Bioinformatics
• Image Analysis
• Computational Biology

Background:

• DNA microarray images contain spots with varying topological properties.
• Analyzing these shape-size characteristics is crucial for accurate data interpretation.
• Traditional methods may not fully capture the textural nuances within each spot.

Purpose of the Study:

• To apply granulometries from mathematical morphology for analyzing shape-size content in DNA microarray spots.
• To develop a method for classifying spots based on their topological and textural features.
• To enable comparison of classified spots with hybridization levels.

Main Methods:

• Mathematical morphology, specifically granulometries and morphological multiscale openings, were employed.
• Analysis involved subtracting successively opened microarrays to identify protrusions.
• Spot-wise probability distributions of protrusions were computed using a grid-based approach.
• Mean size and texture indices were estimated for each spot.

Main Results:

• A method was developed to quantify shape-size distributions within individual microarray spots.
• Spots were successfully classified into 'bright' and 'dull' categories using pattern spectrum and shape-size complexity.
• The approach provides detailed spot-wise textural and size information.

Conclusions:

• Granulometries offer a robust method for characterizing topological variations in DNA microarray spots.
• The classification of spots into bright and dull categories aids in the interpretation of hybridization data.
• This image analysis technique enhances the precision of DNA microarray data assessment.