Ensembl genome and variant annotation
27 September 2026
Jorge Batista da Rocha
Ensembl Outreach, EMBL-EBI
Materials
Introduction to Ensembl�
Why do we need genome browsers?
CGGCCTTTGGGCTCCGCCTTCAGCTCAAGACTTAACTTCCCTCCCAGCTGTCCCAGATGACGCCATCTGAAATTTCTTGGAAACACGATCACTTTAACGGAATATTGCTGTTTTGGGGAAGTGTTTTACAGCTGCTGGGCACGCTGTATTTGCCTTACTTAAGCCCCTGGTAATTGCTGTATTCCGAAGACATGCTGATGGGAATTACCAGGCGGCGTTGGTCTCTAACTGGAGCCCTCTGTCCCCACTAGCCACGCGTCACTGGTTAGCGTGATTGAAACTAAATCGTATGAAAATCCTCTTCTCTAGTCGCACTAGCCACGTTTCGAGTGCTTAATGTGGCTAGTGGCACCGGTTTGGACAGCACAGCTGTAAAATGTTCCCATCCTCACAGTAAGCTGTTACCGTTCCAGGAGATGGGACTGAATTAGAATTCAAACAAATTTTCCAGCGCTTCTGAGTTTTACCTCAGTCACATAATAAGGAATGCATCCCTGTGTAAGTGCATTTTGGTCTTCTGTTTTGCAGACTTATTTACCAAGCATTGGAGGAATATCGTAGGTAAAAATGCCTATTGGATCCAAAGAGAGGCCAACATTTTTTGAAATTTTTAAGACACGCTGCAACAAAGCAGGTATTGACAAATTTTATATAACTTTATAAATTACACCGAGAAAGTGTTTTCTAAAAAATGCTTGCTAAAAACCCAGTACGTCACAGTGTTGCTTAGAACCATAAACTGTTCCTTATGTGTGTATAAATCCAGTTAACAACATAATCATCGTTTGCAGGTTAACCACATGATAAATATAGAACGTCTAGTGGATAAAGAGGAAACTGGCCCC
Genes, regulatory regions, and variation
CGGCCTTTGGGCTCCGCCTTCAGCTCAAGACTTAACTTCCCTCCCAGCTGTCCCAGATGACGCCATCTGAAATTTCTTGGAAACACGATCACTTTAACGGAATATTGCTGTTTTGGGGAAGTGTTTTACAGCTGCTGGGCACGCTGTATTTGCCTTACTTAAGCCCCTGGTAATTGCTGTATTCCGAAGACATGCTGATGGGAATTACCAGGCGGCGTTGGTCTCTAACTGGAGCCCTCTGTCCCCACTAGCCACGCGTCACTGGTTAGCGTGATTGAAACTAAATCGTATGAAAATCCTCTTCTCTAGTCGCACTAGCCACGTTTCGAGTGCTTAATGTGGCTAGTGGCACCGGTTTGGACAGCACAGCTGTAAAATGTTCCCATCCTCACAGTAAGCTGTTACCGTTCCAGGAGATGGGACTGAATTAGAATTCAAACAAATTTTCCAGCGCTTCTGAGTTTTACCTCAGTCACATAATAAGGAATGCATCCCTGTGTAAGTGCATTTTGGTCTTCTGTTTTGCAGACTTATTTACCAAGCATTGGAGGAATATCGTAGGTAAAAATGCCTATTGGATCCAAAGAGAGGCCAACATTTTTTGAAATTTTTAAGACACGCTGCAACAAAGCAGGTATTGACAAATTTTATATAACTTTATAAATTACACCGAGAAAGTGTTTTCTAAAAAATGCTTGCTAAAAACCCAGTACGTCACAGTGTTGCTTAGAACCATAAACTGTTCCTTATGTGTGTATAAATCCAGTTAACAACATAATCATCGTTTGCAGGTTAACCACATGATAAATATAGAACGTCTAGTGGATAAAGAGGAAACTGGCCCCTTGACTAGCAGTAGGAACAATTACTAACAAATCAGAAGCATTAATGTTACTTTATGGCAGAAGTTGTCCAACTTTTTGGTTTCAGTACTCCTTATACTCTTAAAAATGATCTAGGACCCCCGGAGTGCTTTTGTTTATGTAGCTTACCATATTAGAAATTTAAAACTAAGAATTTAAGGCTGGGCGTGGTGGCTCACGCCTGTAATCCCAGCACTTTGGGAGGCCGAGGTGGGCGGATCACTTGAGGCCAGAAGTTTGAGACCAGCCTGGCCAACATGGTGAAACCCTATCTCTACTAAAAATACAAAAAATGTGCTGCGTGTGGTGGTGCGTGCCTGTAATCCCAGCTACACGGGAGGTGGAGGCAGGAGAATCGCTTGAACCCTGGAGGCAGAGGTTGCAGTGAGCCAAGATCATGCCACTGCACTCTAGCCTGGGCCACATAGCATGACTCTGTCTCAAAACAAACAAACAAACAAAAAACTAAGAATTTAAAGTTAATTTACTTAAAAATAATGAAAGCTAACCCATTGCATATTATCACAACATTCTTAGGAAAAATAACTTTTTGAAAACAAGTGAGTGGAATAGTTTTTACATTTTTGCAGTTCTCTTTAATGTCTGGCTAAATAGAGATAGCTGGATTCACTTATCTGTGTCTAATCTGTTATTTTGGTAGAAGTATGTGAAAAAAAATTAACCTCACGTTGAAAAAAGGAATATTTTAATAGTTTTCAGTTACTTTTTGGTATTTTTCCTTGTACTTTGCATAGATTTTTCAAAGATCTAATAGATATACCATAGGTCTTTCCCATGTCGCAACATCATGCAGTGATTATTTGGAAGATAGTGGTGTTCTGAATTA
Regulatory element
Gene
Variant
`
Ensembl features
Ensembl release 2026-07
Ensembl release cycles
Partial Releases (PRs)�For rapid access: new assemblies, gene sets, variation and regulation data released every 2 weeks.
Integrated Releases (IRs)�For long-term stability: frozen snapshots of selected PRs suitable for reproducible research and citations.
IR
Data frozen at a stable reference point
PR 1
PR 2
PR 3
PR 4
PR 5
PR 𝑁
IRs every 3–4 months
PRs every ~2 weeks
Selected PRs are linked together
Human genome assemblies available in Ensembl
Ensembl Variation�
Where does the data come from?
The Ensembl variation process
Variant import
Linking data
Ensembl analysis
Ensembl variation process
Variant import
Linking data
Ensembl analysis
Variant data is imported from publicly available archives and data repositories (varies between species)
Variation sources
Variation annotation
The Ensembl variation process
Variant import
Linking data
Ensembl analysis
Import “value-added” data
Import variant data
Calculate features
Variant consequences
ATG
Regulatory feature
UTR
Intronic
Coding
(non-synonymous/�missense)
Coding
(synonymous)
Splice site
Upstream
UTR
Downstream
https://www.ensembl.org/help/articles/calculated-consequences
AAAAAA
Variant consequences: Sequence ontology (SO)
https://www.ensembl.org/help/articles/calculated-consequences
Variant consequences: Sequence ontology (SO)
https://www.ensembl.org/help/articles/calculated-consequences
Variant interpretation
Variant interpretation
19
You have discovered large numbers of variants from whole exome or genome sequencing - how do you identify the interesting ones?
Common approaches:
20
Import reference variation data
Import phenotype data
Import frequency data
Apply analysis
Web Site*
GraphQL API*
REST API
Ensembl VEP
The Ensembl variation resource build process
*new features may not include all data yet
Variation data in Ensembl - Genome Browser
Variant consequences
Sequence variants
Variation data in Ensembl - Genome Browser
Variation data in Ensembl - Feature explorer
Variation data in Ensembl - Feature explorer
Feature Explorer - Allele frequencies
Ensembl Variant Effect Predictor
Missense variants: Pathogenicity
Various algorithms score or rank changes in amino acid sequence based on:
These are predictions, not facts! A prediction will never be as good as experimental validation.
Missense variants: Pathogenicity scores
1
0
Likely Deleterious
0
0.564
0.34
1
Likely deleterious
Likely
Benign
Uncertain
AlphaMissense
0
30
100
Likely deleterious
Likely benign
CADD
REVEL
Likely benign
Non-coding variants: Impact scores
50
0
Likely Impact
AlphaGenome Variant Impact (AVI) Scores
Likely benign
Alignments viewer for structural variants
Demonstration
We are going to look at the human gene AGT to find variants in the gene.
We will look at the region of AGT to find variants in the region.
We will look at a variant rs699 to find more information about it.
What can you use Ensembl for?
Ensembl integrates gene annotation, variation, regulation and comparative genomics data. You can:
More information
Publication:
Hunt SE, McLaren W, Gil L, et al. Ensembl variation resources. Database: the Journal of Biological Databases and Curation. 2018 Jan;2018. DOI: 10.1093/database/bay119. PMID: 30576484; PMCID: PMC6310513.
Reach out to us and keep up-to-date
The Ensembl Variant Effect Predictor (VEP)
Ensembl VEP
A tool to predict and annotate the functional consequences of variants.
What can you do with the VEP?
VCF
Data input:
Genes, transcripts affected
Pathogenicity
Frequency data
Protein information
PubMed citations
VEP output:
Ensembl VEP interfaces
Functional data
47
IntAct holds evidence of variant impact on interactions, curated from the literature
Experimental results describing the cellular impact of a variant can help interpret organism level phenotype.
Two databases can be integrated into Ensembl VEP via plugins:
MaveDB holds the results of Multiplex Assays of Variant Effect.
These are highly parallel assays to investigate the effect on cellular phenotypes of all possible variants in specific functional regions.
MaveDB: an open-source platform to distribute and interpret data from multiplexed assays of variant effect. Esposito, D. et al. Genome Biol 20, 223 (2019). https://doi.org/10.1186/s13059-019-1845-6
https://www.ebi.ac.uk/intact/home
Variant citation data
48
Provided in Ensembl VEP caches:
Citations can provide additional evidence or otherwise of phenotype association
Links to Genomenon Mastermind variant associations mined from the literature by HGVS-style descriptions are available via a plugin
Caution:
Citation data will include typos in refSNP names and the ambiguity of version-less sequence descriptions
Ensembl VEP plugins are available to integrate predictive scores from tools including:
49
Deleteriousness predictions
Phenotype and functional information is available from sources including :
50
results of highly parallel assays into cellular variant effect
the function, the cellular location and the associated biological processes of the product of any gene a variant lies within
disease associations
evidence of variant impact on interactions
NHGRI-EBI GWAS Catalog
Phenotype/functional evidence plugins
Variant scoring methods
51
=> Agreement on pathogenicity rating may be due to similarity of method or noise in shared input/training data
Tolerance to change
52
Multiple useful measures exist to predict regions where a change is like to have deleterious effect
Conservation measures - show regions which are highly similar across species. eg. GERP, phyloP
Constraint measures - less population variation seen than expected, eg. LOEUF
Dosage Sensitivity - haploinsufficiency (deletion intolerance) and triplosensitivity (duplication intolerance)
The mutational constraint spectrum quantified from variation in 141,456 humans. Karczewski et al doi: https://10.1038/s41586-020-03174-8
A cross-disorder dosage sensitivity map of the human genome. Collins et al. https://doi.org/10.1016/j.cell.2022.06.036
Demonstration
We have identified 10 variants on chromosome 8 in a individual with neurodegenerative symptoms including dystonia.
We will use the Ensembl VEP to determine:
Exercise
Use training.ensembl.org - try the exercises listed in: “Ensembl VEP”
Use the Ensembl VEP to determine:
Ensembl VEP command line
https://e116.ensembl.org/info/docs/tools/vep/script/index.html
Ensembl VEP command line
Use git to download the ensembl-vep package:
Or download the zipped package file:
$ git clone https://github.com/Ensembl/ensembl-vep.git
$ cd ensembl-vep
$ curl -L -O https://github.com/Ensembl/ensembl-vep/archive/release/116.zip
$ unzip 116.zip
$ cd ensembl-vep-release-116/
Ensembl VEP cache - GRCh38 only
Ensembl VEP plugins
More information
Publications:
Hunt SE, Moore B, Amode RM, et al. Annotating and prioritizing genomic variants using the Ensembl Variant Effect Predictor - A tutorial. Human Mutation. 2022 Aug;43(8):986-997. DOI: 10.1002/humu.24298. PMID: 34816521; PMCID: PMC7613081.
McLaren W, Gil L, Hunt SE, et al. The Ensembl Variant Effect Predictor. Genome Biology. 2016 Jun;17(1):122. DOI: 10.1186/s13059-016-0974-4. PMID: 27268795; PMCID: PMC4893825.
Wrap-up�
What are genome browsers?
Genome browsers allow you to search, browse, retrieve, and analyse genomic data.
What is Ensembl?
Ensembl Variation
Citing Ensembl
Yates AD, Austine-Orimoloye O, Azov AG, et al. Ensembl 2026. Nucleic Acids Research. 2025 Nov;gkaf1239. DOI: 10.1093/nar/gkaf1239. PMID: 41296548.
Where to get help
Explore all help and documentation pages, including a guide to the new Ensembl platform.
Data- and tool-specific help pages.
Find out how the data was generated or how to use the tool.
Get in touch with the Ensembl Helpdesk.
Help & documentation
More info
Contact form
Reach out to us and keep up-to-date
Hosting an Ensembl training course
Browser course�One-day course on the Ensembl browser, aimed at wet-lab scientists.
Train-the-trainer course�One-day course on delivering bioinformatics training and teaching your own Ensembl browser course.
Visit the Ensembl Training website or email us for more information.
We can teach an Ensembl course at any institute for free (plus trainers’ expenses in high-income countries).
Acknowledgements and funding
Funded by the European Union
Horizon 2020
Horizon Europe Programme
We would like to acknowledge the National Genome Research Institute (NHGRI) for funding the following grants supporting the creation of the human pangenome reference: U41HG010972, U01HG010971, U01HG013760, U01HG013755, U01HG013748, U01HG013744, R01HG011274, and the Human Pangenome Reference Consortium (BioProject ID: PRJNA730823) as well as 2U24HG007497-09 which supports the creation of HGSVC resources
Funded by Wellcome
Training materials