Ensembl: A Practical Guide

So you need to look up genes, transcripts, variants, or convert IDs between databases? Welcome to Ensembl — the genome browser that bioinformaticians actually use daily.

What is Ensembl?

Ensembl is a genome browser and database jointly run by the EBI (European Bioinformatics Institute) and the Wellcome Trust Sanger Institute since 1999. Think of it as Google Maps, but for genomes.

What it gives you:

• Gene sets (splice variants, proteins, ncRNAs)
• Comparative genomics (alignments, protein trees, orthologues)
• Variation data (SNPs, InDels, CNVs)
• BioMart for bulk data export
• REST API for programmatic access
• Everything is open source

Ensembl Stable Identifiers

This is the ID system you'll see everywhere. Memorize the prefixes:

Transcript Quality Tiers

Not all transcripts are created equal. Here's the hierarchy:

MANE Select (Gold Standard) 🥇

• Matched Annotation between NCBI and EBI
• Perfectly aligned to GRCh38
• Complete sequence identity with RefSeq
• This is your go-to transcript

Merged (Ensembl/Havana) 🥈

• Automatically annotated + manually curated
• High confidence

CCDS (Consensus CDS)

• Collaborative effort for consistent protein-coding annotations
• Shared between NCBI, EBI, UCSC, and others

Ensembl Protein Coding (Red)

• Automatic annotation based on mRNA/protein evidence
• Good, but not manually verified

Using the Ensembl Browser

Basic Navigation

1. Go to ensembl.org
2. Search by: gene name, Ensembl ID, coordinates, or variant ID (rs number)
3. Gene page shows: location, transcripts, variants, orthologues, etc.

Key Information You Can Find

For any gene (e.g., MYH9):

• Ensembl Gene ID → ENSG00000100345
• Chromosomal coordinates → 22:36,281,270-36,393,331
• Cytogenetic location → 22q12.3
• Strand → Forward (+) or Reverse (-)
• Number of transcripts → and which are protein-coding
• MANE Select transcript → with CCDS and RefSeq cross-references

Viewing Variants

1. Navigate to your gene
2. Go to "Variant table" or zoom into a specific region
3. Filter by: consequence type, clinical significance (ClinVar), etc.
4. Click on any variant (e.g., rs80338828) to see:
   • Alleles and frequencies
   • Consequence (missense, synonymous, etc.)
   • Clinical annotations (ClinVar, OMIM)
   • Population frequencies

BioMart: Bulk Data Queries

BioMart is where Ensembl gets powerful. No programming required — it's a web interface for mining data in bulk.

Access: ensembl.org → BioMart (top menu)

The Three-Step Process

1. DATASET    → Choose species/database (e.g., Human genes GRCh38.p14)
2. FILTERS    → Narrow down what you want (gene list, chromosome, biotype...)
3. ATTRIBUTES → Choose what columns to export (IDs, names, sequences...)

⚠️ Common BioMart Mistakes (And How to Avoid Them)

These will save you hours of frustration. Learn from pain.

Mistake #1: Pasting IDs in the Wrong Filter Field

Why it happens: The "Gene stable ID(s)" field expects Ensembl IDs (ENSG...), not RefSeq IDs.

The fix:

Mistake #2: Checkbox vs Text Input Confusion

Some filter options have both a checkbox AND a text field:

☑ With RefSeq peptide ID(s): Only    ← Checkbox (just filters for genes that HAVE RefSeq IDs)
[________________________]           ← Text field (where you paste YOUR specific IDs)

The mistake: Checking the box but not pasting IDs in the text field.

What happens:

• Checkbox alone = "Give me all genes that have ANY RefSeq ID" (thousands of results)
• Text field = "Give me only genes matching THESE specific RefSeq IDs" (your actual query)

The fix: Always paste your ID list in the text input field, not just check the box.

Mistake #3: Orthologue vs Paralogue Mix-up

The fix:

For cross-species queries (e.g., Ciona → Human):

Attributes → Homologues → Human Orthologues
    ✓ Human gene stable ID
    ✓ Human gene name
    ✓ %id. target Human gene identical to query gene

NOT:

Attributes → Homologues → Paralogues   ← WRONG for cross-species!

Mistake #4: Forgetting to Include Filter Column in Attributes

The scenario: You filter by RefSeq peptide ID, but don't include it in your output attributes.

What happens: You get a list of Ensembl IDs with no way to match them back to your original input!

Wait... which RefSeq ID was CASP1 again? 🤷

The fix: Always include your filter field as an output attribute:

Attributes:
    ✓ Gene stable ID
    ✓ Gene name
    ✓ RefSeq peptide ID    ← Include this for verification!

Now you get:

Much better!

Mistake #5: Wrong Dataset for Cross-Species Queries

The scenario: You want human orthologues of Ciona genes. You select "Human genes" as your dataset.

What happens: You can't input Ciona gene IDs because you're in the Human database!

The fix: Start from the source species:

Dataset: Ciona savignyi genes    ← Start here (your input species)
Filters: Gene stable ID → paste Ciona IDs
Attributes: 
    - Gene stable ID (Ciona)
    - Human orthologue gene ID    ← Get human data as attributes
    - Human gene name

Rule: Dataset = species of your INPUT IDs. Other species come through Homologues attributes.

BioMart Mistakes Cheatsheet

Common BioMart Queries

Query Type 1: ID Conversion

RefSeq → Ensembl + HGNC Symbol

Query Type 2: Finding Orthologues

Find human orthologues of genes from another species

Query Type 3: Variant Export

Get all missense variants for a gene list

Query Type 4: Find Genes with PDB Structures

Count/export genes that have associated 3D structures

Practice Exercises

Exercise 1: SNP Nucleotide Lookup

Q: In Ensembl, consider the SNP variation rs80338826. Which is the DNA nucleotide triplet coding for the wild-type amino acid residue (transcript MYH9-201)?

Exercise 2: RefSeq to Ensembl Conversion

Q: Convert these RefSeq protein IDs to Ensembl Gene IDs and HGNC symbols:

NP_203126, NP_001214, NP_001216, NP_001220
NP_036246, NP_203519, NP_203520, NP_203522

Exercise 3: Cross-Species Orthologue Search

Q: Find human orthologues for these Ciona savignyi genes:

ENSCSAVG00000000002, ENSCSAVG00000000003, ENSCSAVG00000000006
ENSCSAVG00000000007, ENSCSAVG00000000009, ENSCSAVG00000000011

Exercise 4: MYH9 Gene Exploration

Q: For the human MYH9 gene:

1. What's the Ensembl code? How many transcripts? All protein-coding? Forward or reverse strand?
2. What's the MANE Select transcript code? CCDS code? RefSeq codes?
3. Chromosomal coordinates? Cytogenetic location?
4. Zoom to exon 17 (22:36,306,051-36,305,930). Any variants annotated in both ClinVar and OMIM? Check rs80338828.

Quick Reference: BioMart Checklist

□ Selected correct dataset (species of your INPUT IDs)
□ Pasted IDs in the CORRECT filter field (match ID prefix!)
□ Used text input field, not just checkbox
□ Selected Orthologue (not Paralogue) for cross-species queries
□ Included filter column in attributes (for verification)
□ Checked "Unique results only" if needed
□ Tested with small subset before full export

TL;DR

• Ensembl = genome browser + database for genes, transcripts, variants, orthologues
• IDs: ENSG (gene), ENST (transcript), ENSP (protein) — learn to recognize them
• MANE Select = highest quality transcript annotation (use these when possible)
• BioMart = bulk query tool: Dataset → Filters → Attributes → Export

Avoid these mistakes:

1. Don't paste RefSeq/UniProt IDs in "Gene stable ID" field — use EXTERNAL filters
2. Use the text input field, not just checkboxes
3. Orthologue = cross-species, Paralogue = same species
4. Start with the species of your INPUT IDs as your dataset
5. Always include your filter column in output attributes

Now go explore some genomes! 🧬