dc.description.abstract | Graph based DNA sequence assembly is actually used to generate contigs form reads produced by second generation sequencer. However, graph based DNA sequence assembly is very sensitive against sequencing error. The existence of sequencing errors will increase the complexity of graph. Meanwhile, every process of sequencing always produce sequencing errors. This research aims to improve the performance of DNA sequencing error correction based on the spectral alignment by implementing a statistical approach. This approach generate the spectrum of solid tuple by choosing tuples that belong to the highest 75% of the tuple occurencies distribution. Reads containing sequencing errors are corrected using tuple references that belong to the solid tuple spectrum. Evaluation is conducted using Velvet, a DNA assembly software. The results show that our approach can reduce the complexity of graph produced by the previous approach up to 45%. | en |