Sequence Mapping of Combinatorial Libraries on Macro- and Microarrays: Bioinformatic Treatment of Data

Sequence mapping of combinatorial libraries is of interest for evaluation of library diversity and homogeneity as well as for bias detection and analysis. This is particularly useful for library improvement by robotic equalization. Macro- and microarray based experimental procedures adapted for this purpose are described in the previous chapters. This chapter aims to delineate procedures and tricks required for data interpretation. Incorrect interpretation of hybridization signals usually occurs because of the highly parallel data treatment using global and automated discrimination criteria when experimental data by themselves are subject to local perturbations and artifacts that would require individual spot examination. In addition, sequence changes involving a few or single base pairs correspond to very limited differences in probe hybridization properties and thus to a low signal-to-noise ratio. We describe here a method for macroarray analysis that is mainly semi-empirical but involves specific data analysis techniques for self-validation. The advantage is of course simplicity, but the drawback derives from subjective and user-dependent interpretations of relative signal intensities. In contrast, the method for interpretation of microarray data relies on statistical analysis and well-defined procedures and is not only giving robust sequence attribution, but also an evaluation of the reliability of the treatment. This method requires the use of a set of specific Excel� macros that are freely distributed.