Modeling of microarray data with zippering

The ability of oligonucleotide microarrays to measure gene expression has been hindered by an imperfect understanding of the relationship between input RNA concentrations and output signals. We argue that this relationship can be understood based on the underlying statistical mechanics of these devices. We present a model that includes the relevant interactions between the molecules. Our model for the first time accounts for partially zippered probe-target hybrids in a physically realistic manner, and also includes target-target binding in solution. Large segments of the target molecules are not bound to the probes, often in an asymmetric pattern, emphasizing the importance of modeling zippering properly. The resultant fit between the model and training data using optimized parameters is excellent, and it also does well at predicting test data.