Biomathematics
Stanford Genome Technology Center

A collaboration among the SGTC1, clinical researchers at Tel Aviv University2, Stanford Department of Biological Sciences3 and the Hebrew University of Jerusalem4 investigates differential gene expression in lung adenocarcinoma, the most common form of lung cancer. Primary tumor tissue and matched histologically non-malignant lung tissue was collected from 18 patients. RNA was extracted and analyzed using Affymetrix U133A microarrays and protocols. Statistical analysis using false discovery rate methods identified 2065 significantly up- and down-regulated genes. Ingenuity Systems Pathways Analysis software, which implements a hierarchical clustering method, was used to focus on the key networks based on interactions drawn from public databases and ontologies, and curated from the research literature. The primary network identified (limited to 35 nodes) illustrated below includes known and putative tumor suppressors, an angiogenesis inhibitor, and a known drug target for the lung cancer drug gemcitabine. Upregulated genes are in red; downregulated in green. Of the 13 genes most highly overexpressed in tumor tissue in this primary network, 8 are known to have clear links to lung cancer (see References)

Significance

• This pilot study shows computational and statistical analysis that provides a useful starting point to identify relevant genes and pathways for further investigation.
larger version, html file incl. 64 kB graphic
• The integration of gene expression data and molecular pathway information helps focus the search for potential biomarkers, drugs and drug targets. It also generates insight and hypotheses for the underlying molecular interactions.
• Matched normal control tissue samples provide much increased statistical power.

Objectives and Approach

• Examine clinical cancer samples with gene expression assays
• Develop statistical false discovery rate method for paired tumor/normal samples
• Employ Ingenuity Systems Pathways Analysis software to identify networks of genes
• Incorporate clinical and pharmaceutical annotation

Data analysis methods are implemented in the statistical language R and provide input for Ingenuity Pathways software. Subsequent literature search provided confirmation of results.

Accomplishments

• Collected and processed paired tumor-normal samples
• False discovery rate (FDR) statistics identified 2065 genes with significantly altered expression
• Tested and applied Ingenuity Systems Pathways Analysis software
• Identified core network of genes affected in adenocarcinoma
• Confirmed known gene-tumor associations and discovered unknown targets which form part of the core adenocarcinoma network