A new approach for filtering noise from high-density oligonucleotide microarray datasets

Although DNA microarrays are powerful tools for profiling gene expression, the dynamic range and the sheer number of signals produced require efficien t procedures for distinguishing false positive results (noise) from changes in expression that are 'real' (independently reproducible). We have develo ped an approach to filter noise from datasets generated when high density o ligonucleotide-based microarrays are used to compare two distinct RNA popul ations. First, we performed comparisons between chips hybridized with cRNAs prepared from an identical starting RNA population; an 'Increase' or 'Decr ease' call in such a comparison was defined as a false positive. Plotting t he average distribution of these false positive signal intensities across 1 8 such comparisons of nine independent RNA preparations allowed us to devel op a series of noise-filtering lookup tables (LUTs). Using a database of 70 separate chip-to-chip comparisons between distinct RNA preparations prepar ed by different workers at different sites and at different times, we show that the LUTs can be used to predict the likelihood that a given transcript called Increased or Decreased in one comparison will again be called Incre ased or Decreased in a replicate comparison. Evidence is presented that thi s LUT-based scoring system provides greater predictive value for reproducib le microarray results than imposition of arbitrary fold-change thresholds a nd accurately predicts which microarray-identified changes will be validate d by Independent assays such as quantitative real-time PCR.