Improved elucidation of biological processes linked to diabetic nephropathy by single probe-based microarray data analysis.

Diabetic nephropathy (DN) is a complex and chronic metabolic
disease that evolves into a progressive fibrosing renal disorder.
Effective transcriptomic profiling of slowly evolving disease
processes such as DN can be problematic. The changes that occur are
often subtle and can escape detection by conventional
oligonucleotide DNA array analyses. We examined microdissected
human renal tissue with or without DN using Affymetrix
oligonucleotide microarrays (HG-U133A) by standard Robust
Multi-array Analysis (RMA). Subsequent gene ontology analysis by
Database for Annotation, Visualization and Integrated Discovery
(DAVID) showed limited detection of biological processes previously
identified as central mechanisms in the development of DN (e.g.
inflammation and angiogenesis). This apparent lack of sensitivity
may be associated with the gene-oriented averaging of
oligonucleotide probe signals, as this includes signals from
cross-hybridizing probes and gene annotation that is based on out
of date genomic data. We then examined the same CEL file data using
a different methodology to determine how well it could correlate
transcriptomic data with observed biology. ChipInspector (CI) is
based on single probe analysis and de novo gene annotation that
bypasses probe set definitions. Both methods, RMA and CI, used at
default settings yielded comparable numbers of differentially
regulated genes. However, when verified by RT-PCR, the single probe
based analysis demonstrated reduced background noise with enhanced
sensitivity and fewer false positives. Using a single probe based
analysis approach with de novo gene annotation allowed an improved
representation of the biological processes linked to the
development and progression of DN. The improved analysis was
exemplified by the detection of Wnt signaling pathway activation in
DN, a process not previously reported to be involved in this
disease.