Design of DNA sequence alignment accelerated system using 2-dimensional array and custom instruction on FPGA / Nur Dalilah Ahmad Sabri
Nowadays, the (deoxyribonucleic acid) DNA sequence database has been increased linearly with the time taken to comparing with the search of DNA sequence alignment system. Hence, the requirements of the computational method for comparing DNA sequences of the sequence alignment area are in high demand...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2018
|
| Online Access: | https://ir.uitm.edu.my/id/eprint/99122/1/99122.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Nowadays, the (deoxyribonucleic acid) DNA sequence database has been increased linearly with the time taken to comparing with the search of DNA sequence alignment system. Hence, the requirements of the computational method for comparing DNA sequences of the sequence alignment area are in high demand. Thus, this research concentrates on the optimization techniques based on Custom Instruction (CI) and Rectangular Dimensional array (2D) that proposed on Smith-Waterman (SW) algorithm. The proposed techniques are used to improve the similarity searching of the DNA sequence alignment system. The system performance is often degraded due to the suffer issues in terms of the time and sensitivity. Other than that, an implementation of a platform called Field Programmable Gate Array (FPGA) has been used in optimizing the 2D Array to perform accelerate sequence alignment activities. The proposed 2D Array on SW algorithm has been developed and designed by using ideas over previous work of CI on SW algorithm. As the result, when comparing the DNA database within 2x2 to 64x64 base pairs with the 2D Array system on FPGA, the acceleration in hardware version gives improvement in the result performance. The highest percentage speedup of the optimization core can goes up to 52% when the proposed technique is applied in aligning two based pair DNA sequences and requires only one bit data storage over matrix cell. In conclusion, the proposed technique shows the result compared to the other algorithms and the results is consistent with the expected theoretical result analysis. Ultimately, the performance time taken to complete the system is reduced proportionally from 1.09us till 0.1617s where it saves the time and aligns the two DNA sequences with 32 sequences in one clock cycle. |
|---|