mySORT

mySORT estimates the relative proportions of twenty-one immune cell subclasses from human tissue transcriptomes profiled by microarray to characterize the tumor microenvironment for cancer immunotherapy and precision medicine.

Key Features:

• Deconvolution algorithm: Two-step deconvolution that performs gene feature selection followed by β-Support Vector Regression (β-SVR) to resolve the relative proportions of twenty-one immune cell subclasses from microarray-profiled human tissue transcriptomes.
• Visualization capabilities: Generation of 2D and 3D plots to compare and visualize immune cell composition across samples.
• Validation on diverse data sets: Performance validated using microarray-based datasets and synthetic pseudo-bulk expression data derived from single-cell transcriptomic datasets of melanoma and head and neck cancer patients.

Scientific Applications:

• Tumor microenvironment analysis: Quantification of immune cell composition to characterize immune contexture within tumors.
• Cancer immunotherapy research: Assessment of immune infiltrates to inform studies of response and mechanism in immunotherapy.
• Biomarker and therapeutic target identification: Support for identifying immune-related biomarkers and potential therapeutic targets for precision medicine applications.

Methodology:

Deconvolution of gene expression via gene feature selection followed by β-Support Vector Regression (β-SVR) to construct a model that predicts immune cell proportions from transcriptomes.