Single object profiles regression analysis (SOPRA): a novel method for analyzing high-content cell-based screens.

Area of Science:

• Cell biology
• Bioinformatics
• Computational biology

Background:

• High-content screening (HCS) generates complex cellular data, often analyzed using averaged values, leading to increased false positives.
• There is a need for improved methods to reproducibly predict hits by identifying significantly altered cell populations.

Purpose of the Study:

• To develop and present SOPRA, a novel workflow for analyzing image-based HCS data.
• To enable reproducible hit prediction by identifying significantly altered cell populations.

Main Methods:

• SOPRA utilizes regression analysis on non-averaged object features from cell populations.
• Data undergoes plate-wise normalization, binning into frequency distribution profiles (histograms), and control-based normalization.
• Analysis is performed using the Bioconductor R-package maSigPro to identify statistically significant altered frequency distributions.

Main Results:

• SOPRA successfully identifies statistically significant normalized frequency distribution profiles.
• The workflow can be applied to hundreds of samples across various cell features.
• Significantly changed profiles can be visualized in heatmaps to identify similar phenotypes and reduce false positives.

Conclusions:

• SOPRA is a novel analysis workflow for detecting statistically significant normalized frequency distribution profiles in high-throughput RNAi screens.
• The workflow was validated using a cell cycle progression screen, identifying profiles for siRNA and chemical inhibitors.
• SOPRA software is freely available on Github.