Describe biological databases (GenBank, UniProt, PDB)
WHY do biological databases exist?
WHY (the problem they solve):
- A single sequencing run can produce billions of bases — no human can eyeball this.
- Discoveries are useless if not shared and comparable. Databases give a common reference so lab A in Japan and lab B in Brazil talk about the same gene.
- They enable comparison (BLAST a new sequence against everything ever deposited) and reuse (build on prior work instead of re-sequencing).
Primary vs Secondary databases — steel-manning the distinction:
- Primary (archival): stores raw experimental submissions as-is (GenBank, PDB). Nothing is "cleaned."
- Secondary (curated): processed/annotated knowledge derived from primaries (UniProt/Swiss-Prot). Humans add function, quality checks.
The three databases
1. GenBank — nucleotide sequences
- WHAT it holds: raw submitted DNA/RNA sequences + metadata.
- HOW you access: through NCBI; search tool = Entrez; similarity search tool = BLAST (Basic Local Alignment Search Tool).
- ID example: an accession like
NM_000518(human beta-globin mRNA). - Format: FASTA (sequence only) or GenBank flat-file (sequence + rich annotation).
2. UniProt — protein sequences & function
- WHAT it holds: protein sequence + function, domains, PTMs, subcellular location, disease links, cross-references.
- WHY curation matters: two databases for a quality vs coverage trade-off — Swiss-Prot is small but trustworthy; TrEMBL is huge but tentative.
- ID example: accession
P68871(human hemoglobin subunit beta).
3. PDB — 3D structures
- WHAT it holds: atomic (x, y, z) coordinates of every atom, plus the method used.
- HOW structures are solved: X-ray crystallography, NMR spectroscopy, cryo-EM.
- ID example: a 4-character code like
1HHO(human oxyhemoglobin). (This 4-char ID is a classic exam giveaway.) - File formats: legacy
.pdb, modern.cif(mmCIF), which handles very large structures. - Quality metric: resolution in Ångströms (Å) — lower is better: 1.5 Å is sharp, 3.5 Å is blurry.
How they connect (central dogma view)
| Feature | GenBank | UniProt | PDB |
|---|---|---|---|
| Data type | Nucleotide (DNA/RNA) | Protein sequence + function | 3D atomic structure |
| Class | Primary | Secondary (curated) + auto | Primary |
| Managed by | NCBI (INSDC) | UniProt Consortium | wwPDB |
| Search/tool | Entrez, BLAST | UniProt search, BLAST | RCSB search |
| ID example | NM_000518 | P68871 | 1HHO |
| Key metric/split | annotation | Swiss-Prot / TrEMBL | resolution (Å) |
Worked reasoning examples
Active Recall
Recall Cover and test yourself
- Which DB stores DNA/RNA? → GenBank
- Which DB stores 3D coordinates? → PDB
- Which DB is the curated protein resource? → UniProt
- GenBank's two international partners? → ENA (Europe), DDBJ (Japan)
- UniProt reviewed vs unreviewed? → Swiss-Prot (manual) vs TrEMBL (auto)
- Is lower or higher resolution (Å) better? → Lower Å = better
- PDB ID format? → 4 characters (e.g., 1HHO)
- Primary vs secondary database? → primary = raw archive; secondary = curated/derived
Recall Feynman: explain to a 12-year-old
Imagine three giant online libraries. The first (GenBank) keeps the spelling of genes — the A, T, G, C letters. The second (UniProt) keeps what proteins do — like a job description for each protein. The third (PDB) keeps 3D LEGO models of what proteins actually look like when folded up. Scientists all over the world put their findings in these libraries so nobody has to start from zero — and the libraries are linked, so one click takes you from the gene's letters, to its job, to its shape.
Flashcards
GenBank stores what type of data?
Who maintains GenBank?
GenBank's international partner databases?
Tool to search similar sequences in GenBank?
UniProt stores what type of data?
What are the two subsets of UniProtKB?
Which UniProt subset is manually reviewed?
PDB stores what type of data?
Three methods to determine structures in PDB?
PDB identifier format?
In crystallography, is lower or higher resolution better?
Primary vs secondary database?
Which consortium manages the PDB?
Central-dogma mapping of the 3 databases?
Modern PDB file format for large structures?
Connections
- Central Dogma of Molecular Biology — the DNA→protein→structure logic each DB captures
- Sequence Alignment & BLAST — the tool that makes GenBank/UniProt searchable
- Protein Structure Determination — X-ray, NMR, cryo-EM behind PDB
- FASTA and File Formats — how sequences are stored
- Protein Folding — why 3D structure (PDB) differs from sequence (UniProt)
- Genome Annotation — turning GenBank raw sequence into UniProt function
Concept Map
Hinglish (regional understanding)
Intuition Hinglish mein samjho
Socho teen badi online libraries hain. GenBank DNA/RNA sequences (A, T, G, C ke letters) store karta hai — yeh ek primary database hai jise NCBI chalata hai, aur iske partners Europe mein ENA aur Japan mein DDBJ hain. Jab aap koi naya gene sequence karte ho, toh use GenBank mein deposit karke ek accession ID (jaise NM_000518) milta hai, aur BLAST tool se aap milte-julte sequences dhoondh sakte ho.
UniProt protein ki duniya hai — yeh protein ka sequence aur uska kaam (function), domains, disease links sab batata hai. Iske do hisse hain: Swiss-Prot jo insaan ne haath se check kiya hai (bharosemand), aur TrEMBL jo machine ne automatic banaya hai (bada lekin abhi unverified). Yaad rakho: "Swiss watch hand-made, TrEMBL trembles kyunki machine ne banaya."
PDB protein ke asli 3D shape ke atomic coordinates (x, y, z) rakhta hai, jo X-ray crystallography, NMR ya cryo-EM se nikalte hain. Iski ID sirf 4 characters ki hoti hai (jaise 1HHO), aur quality ka paimana resolution (Angstrom mein) hai — yahaan chhota number achha hota hai (1.8 Å > 3.2 Å quality mein). Ek hi hemoglobin beta-chain teeno databases mein alag ID ke saath milega — yeh central dogma (DNA → protein → structure) ko exactly follow karta hai, aur cross-links se aap ek click mein hop kar sakte ho. Yeh sab isliye zaroori hai kyunki science mein data share aur compare karna hi asli power hai.