DNA Microarrays and Likelihood Ratio Bioinformatic Methods: Discovery of Human Melanocyte Biomarkers

Abstract

In this article, some of the advantages and limitations of DNA microarray technologies for gene expression profiling are summarized. As a model experiment, DermArray DNA microarrays were utilized to identify potential biomarkers of cultured normal human melanocytes in two different experimental comparisons. In the first case, melanocyte RNA was compared with vastly dissimilar non-melanocytic RNA samples of normal skin keratinocytes and fibroblasts. In the second case, melanocyte RNA was compared with a primary cutaneous melanoma line (MS7) and a metastatic melanoma cell line (SKMel-28). The alternative approaches provide dramatically different lists of 'normal melanocyte' biomarkers. The most robust biomarkers were identified using principal component analysis bioinformatic methods related to likelihood ratios. Only three of 25 robust biomarkers in the melanocyte-proximal study (i.e. melanocytes vs. melanoma cells) were coincidentally identified in the melanocyte-distal study (i.e. melanocytes vs. non-melanocytic cells). Selected up-regulated biomarkers of melanocytes (i.e. TRP-1, melan-A/MART-1, silver/Pmel17, and nidogen-2) were validated by qRT-PCR. Some of the melanocytic biomarkers identified here may be useful in molecular diagnostics, as potential molecular targets for drug discovery, and for understanding the biochemistry of melanocytic cells.

Global profiling of genes modified by endoplasmic reticulum stress in pancreatic beta cells reveals the early degradation of insulin mRNAs

“…There was excellent agreement between the APOCHIP and real-time RT-PCR regarding qualitative changes in gene expression, but the observed quantitative changes were 1.5-2.0-fold lower in the APOCHIP as compared with real-time RT-PCR. Similar findings have been reported in other array systems [29], and may be an inherent limitation of this system. The expression of the genes encoding the pro-apoptotic transcription factors ATF3 and Chop returned to basal levels after the 3 h recovery period, while expression of the genes encoding the key ER chaperones Bip and Grp94 remained elevated (ESM Table 1).…”

Induction of nuclear factor-κB and its downstream genes by TNF-α and IL-1β has a pro-apoptotic role in pancreatic beta cells

“…The quantitative changes, however, were lower in the APOCHIP than those observed by real-time RT-PCR (Figs 3 and 4). These differences were reported in previous array analyses [35,37], and may be an innate limitation of the array technology. The expression of additional NF-κB target genes that were not present (A20) or detected (Cd40) in the APOCHIP was also analysed.…”

Analysis of differential gene expression in human melanocytic tumour lesions by custom made oligonucleotide arrays

“…Our differential gene expression data showed minimal overlap with previously described microarray experiments in which also gene expression patterns were studied in cutaneous melanoma progression (Brem et al, 2001;Baldi et al, 2003b;Carr et al, 2003;Dooley et al, 2003;Rumpler et al, 2003;Becker et al, 2004;Nambiar et al, 2004). This could be explained by the fact that design of our custom made oligonucleotide arrays was based on previously found differential gene expression in melanoma cell lines.…”

“…For most microarray studies related to melanocytic tumorigenesis, cell lines or fresh tumour cells that were cultured for some passages were used to examine differential gene expression (Baldi et al, 2003a;Dooley et al, 2003;Rumpler et al, 2003;Hoek et al, 2004). However, usage of these cells is not ideal, as culturing conditions can influence the genetic expression.…”