Background: Several preprocessing methods are available for the analysis of Affymetrix Genechips arrays. The most\npopular algorithms analyze the measured fluorescence intensities with statistical methods. Here we focus on a\nnovel algorithm, AffyILM, available from Bioconductor, which relies on inputs from hybridization thermodynamics\nand uses an extended Langmuir isotherm model to compute transcript concentrations. These concentrations are\nthen employed in the statistical analysis. We compared the performance of AffyILM and other traditional methods\nboth in the old and in the newest generation of GeneChips.\nResults: Tissue mixture and Latin Square datasets (provided by Affymetrix) were used to assess the performances\nof the differential expression analysis depending on the preprocessing strategy. A correlation analysis conducted\non the tissue mixture data reveals that the median-polish algorithm allows to best summarize AffyILM\nconcentrations computed at the probe-level. Those correlation results are equivalent to the best correlations\nobserved using popular preprocessing methods relying on intensity values. The performances of each tested\npreprocessing algorithm were quantified using the Latin Square HG-U133A dataset, thanks to the comparison of\ndifferential analysis results with the list of spiked genes. The figures of merit generated illustrates that the\nperformances associated to AffyILM(medianpolish), inferred from the present statistical analysis, are comparable to\nthe best performing strategies previously reported.\nConclusions: Converting probe intensities to estimates of target concentrations prior to the statistical analysis,\nAffyILM(medianpolish) is one of the best performing strategy currently available. Using hybridization theory, probelevel\nestimates of target concentrations should be identically distributed. In the future, a probe-level multivariate\nanalysis of the concentrations should be compared to the univariate analysis of probe-set summarized expression\ndata.
Loading....