Genome assembly with Flye for long reads: what coverage depth is needed for a good assembly?

Question

I'm assembling a bacterial genome (~4.5 Mb) using Oxford Nanopore reads with Flye. I have about 15x coverage right now. The assembly is fragmented (150+ contigs). How much coverage do I actually need for a good assembly, and what parameters should I tune?

Admin · Accepted Answer

For bacterial genomes with Flye and Nanopore reads, you generally want **30–60x coverage** for a good assembly. 15x is too low and explains the fragmentation.

Also important: **read quality and length** matter more than raw coverage. N50 read length >10 kb is ideal.

Try these Flye parameters:

```bash
flye 
  --nano-hq reads.fastq.gz 
  --genome-size 4.5m 
  --out-dir assembly_output 
  --threads 16 
  --min-overlap 3000 
  --iterations 3
```

Use `--nano-hq` for R10.4+ reads (>Q20), `--nano-raw` for older chemistry. After assembly, polish with Medaka:

```bash
medaka_consensus -i reads.fastq.gz -d assembly.fasta -o medaka_output -t 8
```

With 30x+ coverage and proper polishing, you should get a complete circular chromosome in 1–3 contigs for most bacteria.

Admin · Answer

If you're stuck with 15x coverage, you can try Raven or Miniasm as alternatives — they sometimes perform better at low coverage:

```bash
raven --threads 16 reads.fastq.gz > assembly.fasta
```

But honestly, sequence more. For Nanopore, a MinION flow cell typically gives 5–15 Gb for R9.4.1, easily enough for 100x+ coverage of a 4.5 Mb genome.