NCBI: A Practical Guide

So you need to search for nucleotide sequences, reference sequences, or gene information? Welcome to NCBI — the American counterpart to Europe's EBI, and home to GenBank, RefSeq, and about 40 other interconnected databases.

What is NCBI?

National Center for Biotechnology Information — created in 1988 as part of the National Library of Medicine (NLM) at NIH, Bethesda, Maryland.

What it gives you:

GenBank (primary nucleotide sequences)
RefSeq (curated reference sequences)
Gene database (gene-centric information)
PubMed (literature)
dbSNP, ClinVar, OMIM (variants & clinical)
BLAST (sequence alignment)
And ~40 more databases, all cross-linked

ℹ️

Global Search

Search any term (e.g., "HBB") from the NCBI homepage and it returns results across ALL databases — Literature, Genes, Proteins, Genomes, Genetics, Chemicals. Then drill down into the specific database you need.

The Three Main Sequence Databases

Database	What it is	Key Point
Nucleotide	Collection from GenBank, RefSeq, TPA, PDB	Primary entry point for sequences
GenBank	Primary archive — anyone can submit	Raw data, may have duplicates/contradictions
RefSeq	Curated, non-redundant reference sequences	Clean, reviewed, NCBI-maintained

GenBank vs RefSeq: Know the Difference

This is crucial — they serve different purposes:

Aspect	GenBank	RefSeq
Curation	Not curated	Curated by NCBI
Who submits	Authors/labs	NCBI creates from existing data
Who revises	Only original author	NCBI updates continuously
Redundancy	Multiple records for same locus	Single record per molecule
Consistency	Records can contradict each other	Consistent, reviewed
Scope	Any species	Model organisms mainly
Data sharing	Shared via INSDC	NCBI exclusive
Analogy	Primary literature	Review articles

💡

When to Use Which?

GenBank: When you need all available sequences, including rare species or unpublished data.
RefSeq: When you need a reliable, canonical reference sequence for analysis.

INSDC: The Global Sequence Collaboration

GenBank doesn't exist in isolation. Since 2005, three databases synchronize daily:

        DDBJ (Japan)
           ↓
    ← → INSDC ← →
           ↓
NCBI/GenBank    ENA/EBI (Europe)
    (USA)

Submit to one, it appears in all three. This is why you sometimes see the same sequence with different accession prefixes.

Understanding Accession Numbers

GenBank Accessions

The LOCUS line tells you a lot:

LOCUS       SCU49845    5028 bp    DNA    PLN    21-JUN-1999
            ↑           ↑          ↑      ↑      ↑
         Name        Length     Type  Division  Date

GenBank Divisions (the 3-letter code):

Code	Division
PRI	Primate sequences
ROD	Rodent sequences
MAM	Other mammalian
VRT	Other vertebrate
INV	Invertebrate
PLN	Plant, fungal, algal
BCT	Bacterial
VRL	Viral
PHG	Bacteriophage
SYN	Synthetic

Query by division: gbdiv_pln[Properties]

RefSeq Accession Prefixes

This is important — the prefix tells you exactly what type of sequence it is:

Prefix	Type	Curation Level
NM_	mRNA	Curated ✓
NP_	Protein	Curated ✓
NR_	Non-coding RNA	Curated ✓
XM_	mRNA	Predicted (computational)
XP_	Protein	Predicted (computational)
XR_	Non-coding RNA	Predicted (computational)
NG_	Genomic region	Reference
NC_	Chromosome	Complete
NT_	Contig	Assembly
NW_	WGS Supercontig	Assembly

⚠️

N vs X Prefix

NM_, NP_ = Curated, experimentally supported
XM_, XP_ = Predicted by algorithms, not yet reviewed

For reliable analyses, prefer N* prefixes when available!

RefSeq Status Codes

Status	Meaning	Reliability
REVIEWED	Reviewed by NCBI staff, literature-backed	⭐⭐⭐ Highest
VALIDATED	Initial review done, preferred sequence	⭐⭐ High
PROVISIONAL	Not yet reviewed, gene association established	⭐ Medium
PREDICTED	Computational prediction, some aspects predicted	⭐ Medium
INFERRED	Predicted, partially supported by homology	Low
MODEL	Automatic pipeline, no individual review	Lowest

NCBI Search Syntax

This is where it gets powerful. NCBI uses field tags in square brackets.

Basic Syntax

search_term[Field Tag]

Boolean operators must be UPPERCASE: AND, OR, NOT

Common Field Tags

Field Tag	What it searches	Example
`[Title]`	Definition line	`glyceraldehyde 3 phosphate dehydrogenase[Title]`
`[Organism]`	NCBI taxonomy	`mouse[Organism]`, `"Homo sapiens"[Organism]`
`[Properties]`	Molecule type, source, etc.	`biomol mrna[Properties]`
`[Filter]`	Subsets of data	`nucleotide omim[Filter]`
`[Gene Name]`	Gene symbol	`BRCA1[Gene Name]`
`[EC/RN Number]`	Enzyme Commission number	`2.1.1.1[EC/RN Number]`
`[Accession]`	Accession number	`NM_001234[Accession]`

Useful Properties Field Terms

Molecule Type

biomol_mrna[Properties]
biomol_genomic[Properties]
biomol_rrna[Properties]

GenBank Division

gbdiv_pri[Properties]    (primates)
gbdiv_rod[Properties]    (rodents)
gbdiv_est[Properties]    (ESTs)
gbdiv_htg[Properties]    (high throughput genomic)

Gene Location

gene_in_mitochondrion[Properties]
gene_in_chloroplast[Properties]
gene_in_genomic[Properties]

Source Database

srcdb_refseq[Properties]           (any RefSeq)
srcdb_refseq_reviewed[Properties]  (reviewed RefSeq only)
srcdb_refseq_validated[Properties] (validated RefSeq only)
srcdb_pdb[Properties]
srcdb_swiss_prot[Properties]

Gene Database Search

The Gene database is the best starting point for gene-specific searches. It integrates information from multiple sources: nomenclature, RefSeqs, maps, pathways, variations, phenotypes.

Gene-Specific Field Tags

Find genes by...	Search syntax
Free text	`human muscular dystrophy`
Gene symbol	`BRCA1[sym]`
Organism	`human[Organism]`
Chromosome	`Y[CHR] AND human[ORGN]`
Gene Ontology term	`"cell adhesion"[GO]` or `10030[GO]`
EC number	`1.9.3.1[EC]`
PubMed ID	`11331580[PMID]`
Accession	`M11313[accn]`

Gene Properties

genetype protein coding[Properties]
genetype pseudo[Properties]
has transcript variants[Properties]
srcdb refseq reviewed[Properties]
feattype regulatory[Properties]

Gene Filters

gene clinvar[Filter]      (has ClinVar entries)
gene omim[Filter]         (has OMIM entries)
gene structure[Filter]    (has 3D structure)
gene type noncoding[Filter]
gene type pseudo[Filter]
src genomic[Filter]
src organelle[Filter]

Building Complex Queries

Query Structure

term1[Field] AND term2[Field] AND (term3[Field] OR term4[Field])

⚠️

Boolean Operators Must Be UPPERCASE

AND, OR, NOT — lowercase won't work!

Example Query Walkthrough

Goal: Find all reviewed/validated RefSeq mRNA entries for mouse enzymes with EC 2.1.1.1 or 2.1.1.10

Breaking it down:

Requirement	Query Component
mRNA sequences	`"biomol mrna"[Properties]`
EC 2.1.1.1 OR 2.1.1.10	`(2.1.1.1[EC/RN Number] OR 2.1.1.10[EC/RN Number])`
Mouse	`"mus musculus"[Organism]`
Reviewed OR validated RefSeq	`("srcdb refseq reviewed"[Properties] OR "srcdb refseq validated"[Properties])`

Final query:

"biomol mrna"[Properties] AND (2.1.1.1[EC/RN Number] OR 2.1.1.10[EC/RN Number]) AND "mus musculus"[Organism] AND ("srcdb refseq reviewed"[Properties] OR "srcdb refseq validated"[Properties])

Result: 9 entries

Practice Exercises

Exercise 1: Nucleotide Database Query

Q: In NCBI "Nucleotide", find all entries containing:

mRNA sequences

coding for enzymes with EC Numbers 2.1.1.1 and 2.1.1.10

from Mus musculus

which have been reviewed or validated in RefSeq

How many entries?

Click for answer

Answer: 9 entries (range: 1-10)

Query:

"biomol mrna"[Properties] AND (2.1.1.1[EC/RN Number] OR 2.1.1.10[EC/RN Number]) AND "mus musculus"[Organism] AND ("srcdb refseq reviewed"[Properties] OR "srcdb refseq validated"[Properties])

How to build it:

Requirement	Field Tag
mRNA	`"biomol mrna"[Properties]`
EC numbers (OR)	`(2.1.1.1[EC/RN Number] OR 2.1.1.10[EC/RN Number])`
Mouse	`"mus musculus"[Organism]`
RefSeq quality	`("srcdb refseq reviewed"[Properties] OR "srcdb refseq validated"[Properties])`

Exercise 2: Gene Database Query

Q: In the «Gene» database, look for all genes:

coding for proteins (protein-coding genes)

associated to the GO term "ATP synthase"

whose source is mitochondrial or genomic

annotated in ClinVar OR OMIM

How many entries?

Click for answer

Answer: 32 entries (range: 31-40)

Query:

"genetype protein coding"[Properties] AND "atp synthase"[Gene Ontology] AND ("source mitochondrion"[Properties] OR "source genomic"[Properties]) AND ("gene clinvar"[Filter] OR "gene omim"[Filter])

How to build it:

Requirement	Field Tag
Protein-coding	`"genetype protein coding"[Properties]`
GO term	`"atp synthase"[Gene Ontology]`
Source (OR)	`("source mitochondrion"[Properties] OR "source genomic"[Properties])`
Clinical (OR)	`("gene clinvar"[Filter] OR "gene omim"[Filter])`

Common Query Patterns

Pattern 1: Species + Molecule Type + Quality

"homo sapiens"[Organism] AND biomol mrna[Properties] AND srcdb refseq reviewed[Properties]

Pattern 2: Gene Function + Clinical Relevance

"kinase"[Gene Ontology] AND gene clinvar[Filter] AND human[Organism]

Pattern 3: Chromosome Region + Gene Type

7[CHR] AND human[ORGN] AND genetype protein coding[Properties]

Pattern 4: Multiple EC Numbers

(1.1.1.1[EC/RN Number] OR 1.1.1.2[EC/RN Number] OR 1.1.1.3[EC/RN Number])

⚠️ Common NCBI Search Mistakes

Mistake #1: Lowercase Boolean Operators

❌ biomol mrna[Properties] and mouse[Organism]
✓ biomol mrna[Properties] AND mouse[Organism]

The fix: Always use UPPERCASE AND, OR, NOT

Mistake #2: Missing Quotes Around Multi-Word Terms

❌ mus musculus[Organism]
✓ "mus musculus"[Organism]

❌ biomol mrna[Properties]
✓ "biomol mrna"[Properties]

The fix: Use quotes around phrases with spaces

Mistake #3: Wrong Database for Your Query

You want...	Use this database
Gene information, GO terms, pathways	Gene
Nucleotide sequences	Nucleotide
Protein sequences	Protein
Variants	dbSNP, ClinVar
Literature	PubMed

Mistake #4: Confusing Properties vs Filters

Type	Purpose	Example
Properties	Content-based attributes	`biomol mrna[Properties]`
Filters	Relationships to other databases	`gene clinvar[Filter]`

Rule of thumb:

Properties = what the sequence IS
Filters = what the sequence is LINKED to

Mistake #5: Using GenBank When You Need RefSeq

If you need a reliable reference sequence for analysis, don't just search Nucleotide — filter for RefSeq:

srcdb refseq[Properties]

Or for highest quality:

srcdb refseq reviewed[Properties]

Quick Reference: Field Tags Cheatsheet

Nucleotide Database

Purpose	Query
mRNA only	`biomol mrna[Properties]`
Genomic DNA	`biomol genomic[Properties]`
RefSeq only	`srcdb refseq[Properties]`
RefSeq reviewed	`srcdb refseq reviewed[Properties]`
Specific organism	`"Homo sapiens"[Organism]`
EC number	`1.1.1.1[EC/RN Number]`
GenBank division	`gbdiv_pri[Properties]`

Gene Database

Purpose	Query
Protein-coding genes	`genetype protein coding[Properties]`
Pseudogenes	`genetype pseudo[Properties]`
GO term	`"term"[Gene Ontology]`
Has ClinVar	`gene clinvar[Filter]`
Has OMIM	`gene omim[Filter]`
Has structure	`gene structure[Filter]`
Chromosome	`7[CHR]`
Gene symbol	`BRCA1[sym]`

Cytogenetic Location Quick Reference

For the Gene database, understanding cytogenetic notation:

    7  q  3  1  .  2
    ↑  ↑  ↑  ↑     ↑
   Chr Arm Region Band Sub-band
   
   p = short arm (petit)
   q = long arm

Example: CFTR gene is at 7q31.2 = Chromosome 7, long arm, region 3, band 1, sub-band 2

TL;DR

NCBI = US hub for biological databases (GenBank, RefSeq, Gene, PubMed, etc.)
GenBank = primary archive (raw submissions) vs RefSeq = curated reference (cleaned up)
RefSeq prefixes: NM/NP = curated, XM/XP = predicted — prefer N* for reliable analysis
Boolean operators MUST be UPPERCASE: AND, OR, NOT
Use quotes around multi-word terms: "homo sapiens"[Organism]
Gene database = best starting point for gene-centric searches
Properties = what it IS, Filters = what it's LINKED to

Now go query some databases! 🧬

Bioinformatics Forever