Modeling oligonucleotide probes for SNP genotyping assays using an adaptive neuro-fuzzy inference system

Developing hypothesis from biological data is one of the main interests and challenges of modern molecular biology. This is somewhat in contrast to the traditional science in this area, where the hypotheses and rules are discovered by biologists, based on simple, often one-parameter-at-a-time, design of experiments. In complex assays, the simple visual exploration of the data deems insufficient. In other words, the abundance, the multiplexity, the multidimensionality, and the low signal-to-noise ratio of the data mandate radically different hypothesis development methods. In this article, the problem of probe design for a multiplexed genotyping assay is studied. Each designed probe has several characteristics, all functions of the nucleic acid sequence. The sequence information can be somewhat condensed by defining a few key features. The objective of this study is to understand the interaction between these features and the way they affect the quality of the probes, as gauged by their performance in the genotyping assay.