Deconvolution of expression microarray data reveals 131I-induced responses otherwise undetected in thyroid tissue

Area of Science:

• Molecular Biology
• Genomics
• Radiation Biology

Background:

• High-throughput gene expression analysis is vital for identifying radiation biomarkers.
• Mixed cell populations in tissue samples can mask cell-specific responses, hindering biomarker discovery.

Purpose of the Study:

• To assess bias in microarray data from mixed thyroid cell populations.
• To evaluate the impact of deconvolution on identifying radiation-induced transcriptional changes.

Main Methods:

• Microarray data from mouse thyroid tissue exposed to 131I radiation were deconvolved using csSAM and R.
• Cell frequencies of follicular cells and C-cells were used for deconvolution.
• Gene Ontology terms and literature-based signature genes analyzed radiation and thyroid hormone responses.

Main Results:

• Deconvolution increased the detection rate of significantly regulated transcripts, including kallikrein.
• Ionizing radiation (IR) and thyroid hormone (TH)-associated genes were more readily detected post-deconvolution.
• Crucial cellular responses (DNA integrity, stress) were identified in deconvolved data, but not in convoluted data.

Conclusions:

• Deconvolution accurately reproduces previously reported 131I-induced transcriptional trends with higher detection rates.
• This method resolves issues in detecting damage and stress responses in heterogeneous samples.
• Deconvolution optimizes microarray data analysis for biomarker screening in complex biological samples.