Error Correction in DNA Computing: Misclassification and Strand Loss

We present a method of transforming an extract-based DNA computation that is error-prone into one that is relatively error-free. These improvements in error rates are achieved without the supposition of any improvements in the reliability of the underlying laboratory techniques. We assume that only two types of errors are possible: a DNA strand may be incorrectly processed or it may be lost entirely. We show to deal with each of these errors individually and then analyze the tradeoff when both must be optimized simultaneously.