Abstract: Next-generation sequencing has generated a deluge of genomic data that transcends the processing capabilities of traditional CPU-based aligners and sequential hardware architectures, which remain constrained by the memory wall and clock-driven execution loops. This paper presents a novel, fully combinational hardware accelerator for exact BWT-FM index DNA sequence alignment, manifested within a non-sequential logic fabric. By synthesising a hardware-native 16-element Bitonic Sorting Network for BWT construction and unrolling the FM-index lookup into a purely spatial architectural chain, the proposed design effectively eliminates the control-flow overhead and sequential state transitions intrinsic to conventional models. The architecture, implemented in Verilog HDL and verified via RTL simulations, achieves 100% mapping accuracy across heterogeneous query workloads while providing a deterministic, sub-cycle latency profile defined solely by the hardware's propagation delay (T_pd). Performance characterisation on a Cyclone IV FPGA confirms that search resolution is decoupled from clock-cycle iterations, converting iterative search algorithms into a spatial logic path. This study establishes a new benchmark for high-throughput genomic mapping, offering a scalable blueprint for real-time clinical diagnostics and advanced genomic surveillance.

Keywords: Bitonic Sorting Network, Burrows-Wheeler Transform (BWT), DNA sequence alignment, Fully combinational logic, FM-index, Hardware acceleration, Field Programmable Gate Array (FPGA), Near-data processing

Pages: 333-356