Basal cell carcinoma (BCC) is the most common cutaneous malignancy that, like other tumours, possesses a heterogeneous genetic composition. In order to select genes with consistent changes in expression among these tumours, we analysed BCC microarray expression data by using a novel approach, termed correlative analysis of microarrays (CAM). CAM is a nested, non-parametric method designed to qualitatively select candidates based on their individual, similar effects upon an array-wide closeness measure. We applied the CAM method to expression data generated by two-channel cDNA microarray experiments, where 21 BCC and patient-matched normal skin specimens were examined. Fifteen candidate genes were selected, with six overexpressed and nine underexpressed in BCC vs. normal skin. Five of the nine consistently downregulated genes in the tumour samples are involved in mitochondrial function and the oxidative phosphorylation (OXPHOS) pathway. One of these genes was the 7.5-kDa subunit, NADH dehydrogenase (ubiquinone) alpha subcomplex-1 (NDUFA1), an accessory component of OXPHOS complex-I that is essential for respiratory activity. These findings support the hypothesis that irregularities in mitochondrial function are involved in neoplasia. Suppression of NDUFA1 expression could represent a key pathogenic mechanism in the development of BCC.