Utilizing high level synthesis to accelerate pattern detection for positive selection based on linkage disequilibrium

Positive selection is an interesting concept in popula-tion genetics, and with more genetic data available new and bettertools to detect selective sweeps indicative of positive selection areneeded. Selective sweeps are detectable by looking for linkagedisequilibrium (LD) and other statistical metrics. RAiSD usesLD and other statistical measures to detect a selective sweep.However with LD it only considers it a LD match when ther2value is 1, and all other values are floored to 0. This raises thequestion whether the performance of RAiSD could be improvedby choosing ther2threshold somewhere else, for example 0.8.This is implemented, but the increase in accuracy of the codeis not worth the increase in computational time with the testeddata sets. This could be investigated further, but a decrease incompute time is required for this. So an FPGA accelerator isdeveloped to try to decrease the computation time and realisethe potential of the system.