NCBI Genomes

NCBI Genomes is the comprehensive genomic sequence database at the National Center for Biotechnology Information. It serves as a centralized repository for assembled genomes from all domains of life, including bacteria, archaea, eukaryotes, and viruses. This guide is for researchers, bioinformaticians, and students who need to retrieve, evaluate, and use genome assemblies from NCBI for comparative genomics, annotation, evolutionary studies, and clinical or environmental microbiology. You will learn how to navigate the database, choose the right assembly, run a practical retrieval workflow, check data quality, avoid common pitfalls, and interpret results within their inherent limits.

NCBI Bookshelf provides authoritative technical references on NCBI databases, including detailed documentation for the Genomes resource. EMBL-EBI Training offers complementary tutorials on genome data retrieval and analysis across international repositories.

At a Glance

Core Concepts

NCBI Genomes organizes genomic data into assemblies, which are the output of genome sequencing and reconstruction. Each assembly is assigned a unique accession number (e.g., GCF_000001405.40 for human GRCh38.p14). Assemblies are grouped by organism and by taxonomic lineage. The database distinguishes between RefSeq assemblies (curated, non-redundant) and GenBank assemblies (original submissions). RefSeq assemblies are generally more reliable because they undergo additional review by NCBI staff.

The assembly level describes how complete the genome is. A "Complete Genome" assembly has no gaps in the chromosome sequences. "Chromosome" level means each chromosome is represented by one or more records but may contain gaps. "Scaffold" level includes ordered and oriented sequences that are not yet joined into full chromosomes. "Contig" level is the most fragmented with no ordering information. The NCBI Prokaryotic Genome Annotation Pipeline automatically annotates many bacterial and archaeal assemblies, adding gene, protein, and functional information.

The Sequence Read Archive (SRA) holds the raw reads that underlie many assemblies. You can often link from an assembly record to its SRA accession and verify the original sequencing data. For example, a study on rice mutant lines used SRA data to discover genetic variations associated with blast resistance Genome-wide data discovery of genetic variations in an advanced rice mutant line, NMR191, with blast resistance and early maturity traits. This direct connection between raw and assembled data is a strength of the NCBI system.

Decision Points

Before downloading a genome from NCBI, you must decide which assembly to use. Consider these criteria.

Organism and strain. The same species may have dozens of assemblies. Choose the one that matches your target strain or isolate. For clinical or epidemiological work, strain information is critical. A study on Klebsiella pneumoniae used topologically defined genomes to track resistance gene evolution Within-host evolution of tandemly duplicated bla(KPC14) conferring cross-resistance to ceftazidime-avibactam, cefiderocol and aztreonam-avibactam in hypervirulent ST11 Klebsiella pneumoniae. Ensure the strain matches your study population.

Assembly level. If you need gene order or synteny, choose the highest level available (Complete Genome or Chromosome). For gene presence/absence analysis, Scaffold or Contig may be sufficient.

RefSeq vs. GenBank. Prefer RefSeq (accession prefix GCF) for consistency and curation. Use GenBank (prefix GCA) when your organism of interest is only available there.

Annotation status. Check whether the assembly has been annotated. For most prokaryotes, NCBI provides annotation automatically. For eukaryotes, annotation may be limited to major model organisms. You may need to perform your own annotation using tools from Galaxy Training Network or Bioconductor.

Assembly date. Newer assemblies may incorporate advanced sequencing technologies and improved scaffolding. Compare statistics like N50 (the length above which half the sequence resides) and total number of contigs. Higher N50 and fewer contigs generally indicate better continuity.

Intended use. If you are conducting a large comparative study, you need uniform annotation. Consider using RefSeq representatives or the GenBank representative assembly from the NCBI Prokaryotic Genomes initiative. For antimicrobial resistance profiling, as in a recent study of 1,240 Stenotrophomonas maltophilia genomes, you rely on assembly quality to accurately detect resistance genes In silico antimicrobial resistance profiling of 1,240 Stenotrophomonas maltophilia genomes: AMR gene prevalence, temporal evolution, and ST-specific associations.

Practical Workflow

Follow this step-by-step process to obtain a genome assembly from NCBI.

1. Identify your target organism. Go to the NCBI Genome website (https://www.ncbi.nlm.nih.gov/genome/) and search by organism name, taxon ID, or higher taxonomic group. Use the filters to limit by assembly level or RefSeq status.

2. Evaluate the available assemblies. Click on an organism page to see a table of assemblies. Review the columns for assembly level, accession, N50, total length, and annotation date. For your chosen assembly, click the accession to open the assembly summary page.

3. Verify metadata and links. On the assembly summary page, check that the organism and strain match your expectation. Look for links to raw sequencing data in the SRA. Confirm that the assembly is not suppressed or withdrawn. Read the "Assurance" note if present.

4. Download the sequence and annotation files. On the right side of the summary page, click "FTP" to access the assembly directory. You will find files with suffixes such as:

• _genomic.fna.gz = nucleotide sequences of all chromosomes/contigs
• _genomic.gff.gz = annotation in GFF3 format
• _genomic.gbff.gz = GenBank flat file with annotations and features
• _protein.faa.gz = protein sequences
• _cds_from_genomic.fna.gz = coding sequence (CDS) nucleotide sequences

Use wget or curl to download the files. For bulk downloads, use the NCBI datasets command-line tool.

5. Validate the downloaded files. Check file integrity using checksums (MD5 files are often provided). Uncompress and inspect the first few lines with zgrep or head to ensure the file contains sequences and that headers are formatted correctly.

6. Perform initial quality control. Use tools like QUAST to assess assembly contiguity and completeness. Compare your assembly's N50 and genome size to known values for the species. For taxonomic validation, screen for contamination using tools such as CheckM or Kraken.

7. Extract or analyze genomic features. Depending on your research question, you may extract gene sequences, compute GC content, identify orthologs, or map reads. If you need to work with multiple assemblies, consider using a workflow manager like Snakemake or Nextflow. The Galaxy Training Network provides step-by-step tutorials for common tasks.

8. Cite the assembly. In publications, cite the assembly accession and the original sequencing publication if available. NCBI provides a citation note on the assembly page.

Quality Checks

Not all assemblies are equally reliable. Perform these checks before using a genome in downstream analysis.

Check assembly statistics. N50, L50 (number of contigs needed to cover half the genome), and total number of gaps. For a bacterial genome, a high-quality complete assembly should have N50 equal to the genome size and zero gaps. For eukaryotic genomes, N50 of several megabases is typical.

Examine annotation completeness. Use BUSCO (Benchmarking Universal Single Copy Orthologs) to evaluate how many conserved orthologs are present. The tool is available through Bioconductor and Galaxy. A low BUSCO score indicates missing genes or assembly errors.

Look for contamination. Foreign sequences (e.g., vector, host, or symbiont DNA) can confound results. Many NCBI assemblies have been screened for contamination, but you should verify with alignment to a clean reference.

Read the publication. If the assembly is associated with a published paper, read it to understand the sequencing technology, coverage, and assembly method. For example, the complete mitochondrial genomes of two flea species were assembled from high coverage sequencing and validated by phylogenetic placement The first complete mitochondrial genome sequences of Neopsylla specialis dechingensis and Neopsylla stevensi sichuanyunnana, with an assessment of their phylogenetic placement. Such validation adds confidence.

Review the RefSeq category. RefSeq genomes are categorized as "reference genome" (highest quality, usually for well-studied organisms), "representative genome" (one per species), or "na" (not curated). Choose reference or representative for most analyses.

Common Mistakes

Using the wrong strain. Many species have multiple assemblies from different isolates. Using a strain that does not match your experimental or clinical material can mislead findings. Always verify the strain lineage.

Ignoring assembly level. Downloading a contig-level assembly when you need chromosomal context forces you to reorder sequences manually. Select the highest level available for your purpose.

Overlooking annotation differences. The same species may have assemblies with different annotation pipelines. When comparing gene content across strains, use assemblies with the same annotation source or perform a consistent reannotation.

Neglecting contamination. A genome with 5% contamination from another bacterium will produce erroneous phylogenies and resistance gene calls. Always run contamination checks.

Assuming completeness. A "Complete Genome" label does not guarantee that every repetitive region is resolved. Telomeres, centromeres, and ribosomal DNA arrays are often missing or collapsed. Be cautious in analyses involving repeats.

Failing to cite properly. Omitting the assembly accession and sequence publication undermines reproducibility. NCBI provides a readiness reporting system, use the provided citation.

Limits of Interpretation

NCBI Genomes is an incredible resource, but it has boundaries.

Not all organisms are represented. The database is biased toward model organisms, pathogens, and economically important species. Many environmental or rare taxa have no assembly. The human oral microbiome, for instance, has a global landscape of publicly available data, but many species remain uncultured and unassembled Global Landscape of Publicly Available Human Oral Microbiome Data.

Assembly quality varies by taxa. Prokaryotic genomes tend to be high quality due to smaller size and standardized pipelines. Eukaryotic genomes, especially those of non-model organisms, may be highly fragmented with many gaps. A new species description may rely on a draft assembly with limited gene annotation Moraxella tarda sp. nov., a Novel Species Isolated from Nasal Cavity of Cattle with Infectious Bovine Keratoconjunctivitis. Such assemblies are useful for taxonomy but less so for deep functional analysis.

Annotation is not experimental validation. Predicted genes are based on algorithms and may miss or incorrectly assign functions. Even RefSeq annotation can contain errors. Always confirm critical genes with read coverage, expression data, or laboratory experiments.

Assembly does not capture all genomic variation. A single reference assembly cannot represent the pan genome of a species. Structural variants, copy number variations, and plasmids may be absent. Use pangenome approaches when studying diverse populations.

Metagenomes are not yet fully integrated. NCBI Genomes is focused on isolate genomes. Metagenome-assembled genomes (MAGs) are available in separate databases like MGnify or IMG. The SRA houses raw metagenome reads, but assembling them into high-quality genomes remains an active research area.

Data are static but biology is dynamic. An assembly represents one point in time. Some assemblies are never updated. If you need the latest version, check the "Assembly update history" on the summary page.

Frequently Asked Questions

Q1: How do I find the RefSeq representative genome for a species?

A: On the organism page on the NCBI Genome website, look for the row labeled "representative genome" under the RefSeq category. This is the single most curated assembly for that species. You can also use the NCBI Datasets command line tool with the option to download the representative genome.

Q2: Can I download multiple genomes at once?

A: Yes. Use the NCBI Datasets command line tool (datasets download genome) to batch download assemblies by taxon, accession list, or search terms. Alternatively, use the FTP directories and write a script to iterate over accessions. For large scale studies, consider using the NCBI Genome metadata report in tab separated format to plan downloads.

Q3: How do I check if a genome has been screened for contamination?

A: On the assembly summary page, look for a section called "Contamination Screening." Many RefSeq assemblies have a "Cleaned" status. You can also view the "ASM" report which notes any removal of contaminated sequences. If you are unsure, run your own contamination screen with tools like CheckM.

Q4: My assembly lacks annotation. How can I add gene predictions?

A: You can use the Prokaryotic Genome Annotation Pipeline (PGAP) if your assembly is bacterial. For eukaryotic genomes, use tools like AUGUSTUS or MAKER, which are available through Galaxy or Bioconductor. Consider submitting your assembly to NCBI for official annotation, however, that service is typically for original submitters.

References and Further Reading

• NCBI Bookshelf contains the NCBI Handbook, a critical reference for database structure and data policies.
• EMBL-EBI Training offers courses on genome browsing, assembly quality, and comparative genomics.
• Galaxy Training Network provides free tutorials for genome assembly quality assessment and annotation.
• Bioconductor supplies R packages for analyzing genome sequences and annotations.
• NCBI Sequence Read Archive allows direct retrieval of raw sequencing data linked to assemblies.
• Use PubMed articles for domain specific validation, such as this rice mutant study Genome-wide data discovery of genetic variations in an advanced rice mutant line and this Klebsiella resistance tracking Within-host evolution of tandemly duplicated bla(KPC14).
• For AMR profiling methodology, see In silico antimicrobial resistance profiling of 1,240 Stenotrophomonas maltophilia genomes.
• For novel species assembly standards, review Moraxella tarda sp. nov..
• For mitochondrial genome assembly in fleas, see The first complete mitochondrial genome sequences of Neopsylla.
• The human oral microbiome data landscape is covered in Global Landscape of Publicly Available Human Oral Microbiome Data.

Related Articles

• Protein Synthesis
• Incomplete Dominance Gene
• Cell Membrane Function Biology
• Dna Structure
• Protein Structure