phenoDist

phenoDist measures phenotypic distance in image-based high-throughput screening datasets to quantify phenotypic dissimilarities, identify strong phenotypes, and group treatments into functional clusters.

Key Features:

• Phenotypic Dissimilarity Measurement: Computes phenotypic dissimilarity between cell populations using Support Vector Machine (SVM) classification combined with cross-validation.
• Automated Parameter Optimization: Automatically optimizes two analysis parameters to adapt to input data.
• High-Throughput Imaging Support: Designed for automated microscopy datasets that generate large, multidimensional feature sets extracted from individual cells.
• Applicability to RNAi Perturbations: Applied to RNA interference (RNAi) and siRNA perturbations, including kinome RNAi screening datasets.
• Treatment Clustering: Uses measured phenotypic distances to cluster treatments into functional groups.
• Replicate Reproducibility and Separation: Produces results with good replicate reproducibility and effective separation between control groups and treated samples.

Scientific Applications:

• Identify Strong Phenotypes: Distinguishes significant phenotypic changes induced by siRNA treatments.
• Discover Functional Links Between Genes: Reveals potential functional connections between genes using kinome RNAi screening datasets.
• Cluster Treatments into Functional Groups: Organizes treatments by functional similarity based on phenotypic distance measurements.

Methodology:

Computes phenotypic dissimilarity via SVM classification with cross-validation and automatic optimization of two parameters to support clustering and separation of control versus treated samples.