Unveiling the immune microenvironment of complex tissues and tumors in transcriptomics through a deconvolution...

Area of Science:

• Computational biology
• Bioinformatics
• Immunology

Background:

• Accurate analysis of tumor-infiltrating leukocytes is crucial for advancing cancer immunotherapy.
• Current bioinformatics approaches for immune microenvironment analysis are complex and challenging.
• Understanding immune cell composition is vital for predicting treatment response.

Purpose of the Study:

• To develop a streamlined bioinformatics workflow for analyzing immune cell composition from RNA sequencing data.
• To create a user-friendly toolkit (DOCexpress_fastqc) and web application (mySORT) for enhanced immune profiling.
• To provide a more accurate and accessible method for deciphering the tumor immune microenvironment.

Main Methods:

• Developed a two-step workflow integrating DOCexpress_fastqc (hisat2-stringtie pipeline in Galaxy/Docker) for RNA sequencing gene expression profiling.
• Utilized mySORT, a web application employing a deconvolution algorithm, to determine immune cell content across 21 subclasses.
• Validated mySORT using synthetic pseudo-bulk data from single-cell RNA sequencing (scRNA-seq) datasets.

Main Results:

• mySORT demonstrated strong concordance with ground-truth immune cell composition (Pearson's r = 0.871 for melanoma, 0.775 for head and neck cancer).
• mySORT outperformed existing methods like CIBERSORT in accuracy for immune cell deconvolution.
• The toolkit offers enhanced resolution for bulk RNA sequencing data and includes data visualization features.

Conclusions:

• The developed workflow and mySORT tool provide an accurate and efficient method for analyzing immune cell composition in cancer.
• This approach facilitates a deeper understanding of the tumor immune microenvironment, crucial for immunotherapy development.
• The freely available toolkit and web application empower researchers to advance cancer immunotherapy strategies.