A microarray data analysis method to evaluate the impact of contaminants on wild animals

Here we propose a novel microarray data analysis method applicable to evaluation of the chemical effects on wild animals. First, we analyzed correlations between log-transformed hepatic 2,3,7,8-tetrachlorodibenzo-p-dioxin toxic equivalent (TEQ) levels and probe signals detected in wild cormorant liver to screen contaminant-responsive genes. Second, principal component analysis (PCA) was conducted using the screened probes. Third, these probes were divided into two groups based on our PCA result. Finally, we calculated Euclidian distance of signals, which is equivalent to variance of gene expressions, in each probe set, and analyzed the relationship between log-transformed hepatic TEQ levels and Euclidian distances. A probe set whereby the calculated Euclidian distance was positively correlated with TEQ levels, could indicate genes that were directly affected by dioxins or other persistent organic pollutants (POPs), hence they can be used as biomarkers. By contrast, there were a number of probes whereby the Euclidian distance was negatively correlated with TEQ levels. In the latter probe group, the smaller Euclidian distances in highly contaminated individuals could point to changes in physiological activities of wild cormorants. Therefore, our microarray data analysis method will provide new insights into POPs-responsive genes in field-collected samples for toxicogenomics studies.