A regression system for estimation of errors introduced by confocal imaging into gene expression data in situ

Area of Science:

• Confocal microscopy
• Image analysis
• Biophotonics

Background:

• Confocal image accuracy is limited by experimental errors during scanning.
• Averaging multiple scans reduces noise but high PMT settings can cause clipping and data errors.
• Existing error correction methods require all raw scans, which are not always available.

Purpose of the Study:

• To develop a system for predicting and correcting errors in averaged confocal images.
• To enable accurate data extraction from images with increased dynamical range.

Main Methods:

• Developed a regression system trained on images from diverse microscopes and PMT settings.
• The system utilizes all available confocal scans for training and prediction.
• The prediction method is implemented as the freely available software tool CorrectPattern.

Main Results:

• The regression system demonstrated high accuracy in predicting error sizes.
• Applied the system to correct errors in gene expression segmentation data in the FlyEx database.
• Successfully corrected errors in strongly clipped images, improving data quality.

Conclusions:

• A novel regression system and software (CorrectPattern) accurately predict and correct confocal image errors.
• The method allows for obtaining higher dynamical range images.
• Enables extraction of more detailed quantitative information from microscopy data.