Restricting SBH ambiguity via restriction enzymes

Abstract. The expected number of n-base long sequences consistent with a given SBH spectrum grows exponentially with n, which severely limits the potential range of applicability of SBH even in an error-free setting. Restriction enzymes (RE) recognize specific patterns and cut the DNA molecule at all locations of that pattern. The output of a restriction assay is the set of lengths of the resulting fragments. By augmenting the SBH spectrum with the target string’s RE spectrum, we can eliminate much of the ambiguity of SBH. In this paper, we build on [20] to enhance the resolving power of restriction enzymes. We give a hardness result for the SBH+RE problem, and supply improved heuristics for the existing backtracking algorithm. We prove a lower bound on the number restric-tion enzymes required for unique reconstruction, and show experimental results that are not far from this bound.