Abstracts: 

 SBIR-Phase I 



Rapid Genome Analysis on a Workstation with an 

 Associative Coprocessor 



Charles D. Stormon, James Brule, and Hamid Bacha 

 Coherent Research, Inc., Syracuse. NY 13202 

 (315)426-0929 



The goal of this research project is to produce a high-performance, low-cost sequence 

 analysis workstation that will allow genetic researchers to perfomi complex analyses on 

 genome data. Phase I research involves feasibility studies of processing genome data at 

 a workstation with an associative coprocessor. Many of the DNA sequence analy.ses that 

 molecular biologists require involve finding similarities between different molecules or 

 within one molecule. Several different approaches, represented by dozens of programs, 

 have been taken to solve this problem. With increasing sequence length (or larger 

 numbers of fragments) to be analyzed, these methods can quickly become compu- 

 tationally impractical. Recently, parallel processors have become available, and they 

 provide an avenue of speedup for sequence analysis. 



By implementing a simplified computational model in VLSI (very large scale 

 integrated), our associative Coherent Processor^*^ (the AP-DS) provides the same 

 degree of parallel processing as either the Connection Machine™ (Thinking Machines, 

 Inc.) or the DAP^" (Active Memory Technology) at a much lower cost. The research in 

 this project will show the feasibility of processing complex genome queries on a 

 workstation-class machine with attached AP-DS coprocessor. 



108 



