Toward a Practical Data Privacy Scheme for a Distributed Implementation of the Smith-Waterman Genome Sequence Comparison Algorithm

Volunteer distributed computations utilize spare processor cycles of personal computers that are connected to the Internet. The resulting platforms provide computational power previously available only through the use of expensive clusters or supercomputers. However, distributed computations running in untrustworthy environments raise a number of security concerns, including computation integrity and data privacy. This paper introduces a strategy for enhancing data privacy in some distributed volunteer computations, providing an important first step toward a general data privacy solution for these computations. The strategy is used to provide enhanced data privacy for the Smith-Waterman local nucleotide sequence comparison algorithm. Our modified Smith-Waterman algorithm provides reasonable performance, identifying most, and in many cases all, sequence pairs that exhibit statistically significant similarity according to the unmodified algorithm, with reasonable levels of false positives. Moreover the modified algorithm achieves a net decrease in execution time, with no increase in memory requirements. Most importantly, our scheme represents an important first step toward providing data privacy for a practical and important real-world algorithm.