A method is presented for de novo assembly of bacterial genomes using only nanopore sequencing data from the Oxford Nanopore MinION. The approach involves detecting overlaps between nanopore reads, correcting errors using partial order graphs, and assembling the corrected reads with the Celera assembler. This method successfully assembled nanopore reads from Escherichia coli K-12 MG1655 into a single contig of 4.6 Mb, allowing full reconstruction of gene order. The resulting assembly had 98.4% nucleotide identity to the reference genome. The MinION is a portable, single-molecule sequencing instrument that can generate long DNA fragments without amplification, making it suitable for genome assembly. Long reads are essential for assembling genomes due to their ability to span repetitive elements. The accuracy of nanopore reads is a key factor in genome assembly, and recent improvements in base calling have increased read accuracy to 78-85%. This study used high-quality "passing filter" two-direction reads from four MinION runs, resulting in 22,270 2D reads and 133.6 Mb of read data. The reads were processed using DALIGNER to detect overlaps, then aligned using POA with partial order graphs to compute a consensus sequence. The corrected reads were assembled with Celera Assembler, with parameters set to ensure high contiguity. The final assembly consisted of four contigs, with the largest being 4.6 Mb and covering the entire E. coli chromosome. The assembly had 3,949 mismatches and 47,395 indels compared to the reference genome. The study found that nanopore sequencing data can be used to assemble complete bacterial genomes without complementary sequencing technologies. However, further improvements in error modeling are needed to enhance base accuracy. The software pipeline used to generate these assemblies is freely available online. The results demonstrate that long read data from the MinION can provide an accurate view of gene order and synteny in bacterial genomes.A method is presented for de novo assembly of bacterial genomes using only nanopore sequencing data from the Oxford Nanopore MinION. The approach involves detecting overlaps between nanopore reads, correcting errors using partial order graphs, and assembling the corrected reads with the Celera assembler. This method successfully assembled nanopore reads from Escherichia coli K-12 MG1655 into a single contig of 4.6 Mb, allowing full reconstruction of gene order. The resulting assembly had 98.4% nucleotide identity to the reference genome. The MinION is a portable, single-molecule sequencing instrument that can generate long DNA fragments without amplification, making it suitable for genome assembly. Long reads are essential for assembling genomes due to their ability to span repetitive elements. The accuracy of nanopore reads is a key factor in genome assembly, and recent improvements in base calling have increased read accuracy to 78-85%. This study used high-quality "passing filter" two-direction reads from four MinION runs, resulting in 22,270 2D reads and 133.6 Mb of read data. The reads were processed using DALIGNER to detect overlaps, then aligned using POA with partial order graphs to compute a consensus sequence. The corrected reads were assembled with Celera Assembler, with parameters set to ensure high contiguity. The final assembly consisted of four contigs, with the largest being 4.6 Mb and covering the entire E. coli chromosome. The assembly had 3,949 mismatches and 47,395 indels compared to the reference genome. The study found that nanopore sequencing data can be used to assemble complete bacterial genomes without complementary sequencing technologies. However, further improvements in error modeling are needed to enhance base accuracy. The software pipeline used to generate these assemblies is freely available online. The results demonstrate that long read data from the MinION can provide an accurate view of gene order and synteny in bacterial genomes.