The Total DNA Homology Experiment
Abstract:
This report describes briefly an experiment which may help answer fundamental questions in molecular biology. In essence, we plan to compare all known DNA sequences with each other, filtering out close matches for further analysis. The size of the computation, in terms of total operations, makes it one of
the largest attempted for any purpose; the experiment only becomes feasible when massive parallelism is employed. To this end, we have already designed, fabricated, and tested a systolic array for DNA sequence matching. Preliminary benchmarks indicate that it is hundreds of times faster than current
minicomputers. Using a small number of these chips, we will be able to complete the experiment in one year's time.