Large-scale image-based profiling of single-cell phenotypes in arrayed CRISPR-Cas9 gene perturbation screens

Area of Science:

• Cell Biology
• Genomics
• Bioinformatics

Background:

• High-content imaging and computer vision enable detailed single-cell phenotype analysis.
• CRISPR-Cas9 gene editing facilitates targeted genetic modifications.
• Integrating these technologies allows for large-scale gene perturbation studies.

Purpose of the Study:

• To present methods for applying CRISPR-Cas9 in large-scale, image-based gene perturbation experiments.
• To profile genes affecting cellular morphology and nuclear pore complex (NPC) component localization.
• To develop a machine learning approach for scoring gene perturbations in image-based screens.

Main Methods:

• Development of a pipeline for constructing a 2,281-plasmid CRISPR-Cas9 targeting library.
• Application of CRISPR-Cas9 in human tissue culture cells compatible with rapid phenotyping.
• Utilizing a machine-learning method to analyze genetic heterogeneity and score perturbations.
• High-content imaging for multivariate profiling of single-cell phenotypes.

Main Results:

• Successful CRISPR-Cas9-mediated gene perturbation in human cells within a phenotyping-compatible timeframe.
• Construction and profiling of a large-scale arrayed library of sequence-verified CRISPR-Cas9 plasmids.
• Identification of genes influencing cellular morphology and NPC component localization.
• Demonstration of a machine-learning approach for scoring and identifying perturbed cells.

Conclusions:

• CRISPR-Cas9 is suitable for large-scale, image-based gene perturbation screens.
• The developed pipeline and machine learning method enable genome-scale, multivariate gene perturbation profiling.
• This approach facilitates the discovery of genes involved in cellular morphology and NPC function.