In short
- Contiguity metrics such as N50 and contig count show how fragmented an assembly is, not whether it is correct.
- Completeness and contamination, estimated from conserved marker genes with tools such as CheckM, show whether the whole genome is there and whether it comes from one organism.
- A widely used standard defines a high-quality draft genome as more than 90% complete with less than 5% contamination.
- Genome size and GC content should match the expected species. Large differences point to contamination or a misidentified strain.
Contiguity: how fragmented is the assembly?
The most quoted assembly statistic is N50. Sort the contigs from longest to shortest and add up their lengths. N50 is the length of the contig at which the running total reaches half of the assembly. For example, an assembly of contigs of 1,000, 800, 500, 400, 200 and 100 kb totals 3,000 kb. The first two contigs add up to 1,800 kb, which passes the halfway point of 1,500 kb, so N50 is 800 kb.
Other contiguity measures include the number of contigs, the largest contig and the total length. QUAST calculates these and many more, and can evaluate an assembly with or without a reference genome [1]. A higher N50 usually means a more useful assembly, but contiguity is not correctness. A misassembly that joins unrelated sequences raises N50 while making the genome wrong.
Completeness and contamination
To judge whether a genome is complete and pure, CheckM looks for marker genes that are expected once in every genome of a given lineage. Missing markers lower the completeness estimate, and markers found more than once suggest contamination with another organism [2]. When CheckM was applied to public isolate genomes, it found a diverse range of errors [2]. Problems are not limited to new data.
CheckM2 uses machine learning instead of fixed marker sets. It is faster, its database is easier to update, and it remains accurate for novel lineages and reduced genomes [3].
Which thresholds to use
| Metric | What it tells you | Benchmark |
|---|---|---|
| Completeness | How much of the expected genome is present | > 90% for a high-quality draft [4] |
| Contamination | Whether sequences from other organisms are mixed in | < 5% for a high-quality draft [4]; isolates are usually far lower |
| Genome size and GC content | Whether the assembly matches the expected species | Close to published genomes of the species |
| Number of contigs | How fragmented the assembly is | Depends on technology: tens to a few hundred for short reads, one per replicon for a complete genome |
| Coverage | How much data supports the assembly | About 50–100× after trimming for short reads [5] |
The completeness and contamination thresholds come from the MIMAG standard, which was written for genomes assembled from metagenomes and single cells and also includes rRNA and tRNA criteria [4]. For isolate genomes they are a useful minimum. A pure culture should give contamination close to zero, so even a few percent deserves a closer look. Coverage targets are covered in how much sequencing coverage you need.
When an assembly fails QC
- High contamination: the culture may be mixed. Re-streak a single colony, check purity and sequence again.
- Low completeness: often caused by low coverage or poor DNA quality. More data may help.
- Unexpected genome size or GC content: suspect contamination or a different species, and check identity with ANI.
- Hundreds of short contigs: low or uneven coverage, or a repeat-rich genome. Long reads can help. See choosing an assembly strategy.
Assembly QC in BacSeq reports
Assembly QC is part of every BacSeq Analytics package, from Basic upwards. The report lists genome size, GC content, contig count, N50 and coverage, and flags any assembly that doesn't meet quality thresholds before further analysis is reported. Results from a low-quality assembly are clearly marked, so they aren't over-interpreted.
References
- Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013;29:1072–1075. doi:10.1093/bioinformatics/btt086
- Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 2015;25:1043–1055. doi:10.1101/gr.186072.114
- Chklovski A, Parks DH, Woodcroft BJ, Tyson GW. CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning. Nat Methods. 2023;20:1203–1212. doi:10.1038/s41592-023-01940-w
- Bowers RM, Kyrpides NC, Stepanauskas R, et al. Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nat Biotechnol. 2017;35:725–731. doi:10.1038/nbt.3893
- Desai A, Marwah VS, Yadav A, et al. Identification of optimum sequencing depth especially for de novo genome assembly of small genomes using next generation sequencing data. PLoS One. 2013;8(4):e60204. doi:10.1371/journal.pone.0060204