ATG5

Chr 6AR

autophagy related 5

Also known as: APG5, APG5-LIKE, APG5L, ASP, SCAR25, hAPG5

NCBI Gene: 9474 ENSG00000057663 UniProt: Q9H1Y0 OMIM: 604261

ATG5 encodes a protein that functions as an E1-like activating enzyme in autophagy, essential for autophagic vesicle formation, mitochondrial quality control, lymphocyte development, and axon maintenance. Mutations cause autosomal recessive spinocerebellar ataxia (SCAR25). The gene is highly constrained against loss-of-function variants (pLI 0.98, LOEUF 0.30), indicating that complete loss of ATG5 function is likely detrimental.

OMIM ResearchSummary from RefSeq, OMIM, UniProt

The protein encoded by this gene, in combination with autophagy protein 12, functions as an E1-like activating enzyme in a ubiquitin-like conjugating system. The encoded protein is involved in several cellular processes, including autophagic vesicle formation, mitochondrial quality control after oxidative damage, negative regulation of the innate antiviral immune response, lymphocyte development and proliferation, MHC II antigen presentation, adipocyte differentiation, and apoptosis. Several transcript variants encoding different protein isoforms have been found for this gene. [provided by RefSeq, Sep 2015]

OMIM ResearchGenerating clinical summary…

Primary Disease Associations & Inheritance

?Spinocerebellar ataxia, autosomal recessive 25MIM #617584

Clinical highlights

Gene–disease validity (ClinGen)

Limited · spinocerebellar ataxia, autosomal recessive 25 · AR · ClinGen 2025

not for standalone diagnostic reporting

Interpreting a novel variant

This gene is strongly intolerant of loss-of-function variation in the population, so LoF variants warrant close attention. No curated mechanism annotation is available — see the mechanism card for the computational prediction and its caveats.Based on population constraint only.

Active trials

793

Pubs (1 yr)

—

P/LP submissions

—

P/LP missense

0.29

LOEUF· LoF intol.

LOF

Mechanism· predicted

Population Genetics & Constraint

gnomAD v4 — loss-of-function & missense intolerance

gnomAD

LoF intolerant — likely haploinsufficient

LoF Constraint?

0.29LOEUF

pLI 0.978

Z-score 3.48

OE 0.06 (0.02–0.29)

Highly constrained

Highly LoF-intolerant (top ~10% of genes)

Missense Constraint?

1.82Z-score

OE missense 0.57 (0.48–0.69)

81 obs / 142.0 exp

Tolerant

Mild missense constraint

Observed / Expected Ratios?

LoF OE?0.06 (0.02–0.29)

0≤0.351.4

Missense OE?0.57 (0.48–0.69)

0≤0.61.4

Synonymous OE?0.83

0≤1.21.6

LoF obs/exp: 1 / 16.1Missense obs/exp: 81 / 142.0Syn Z: 0.89

Background carrier burden — how many people carry a rare variant somewhere in ATG5?

gnomAD v4 · ~731k individuals · PASS only

Rare means allele frequency under

Rare missense

1 in 460

224 rare PASS variants

Rare truncating (LoF)

1 in 6,300

38 rare PASS variants

Rare synonymous (reference)

1 in 990

80 rare PASS variants

Truncating variants are far rarer here than missense — that gap, in human units, is what constraint means for ATG5. A background rate for each class, not the frequency of any particular variant; thin-coverage exons undercount.

By genetic ancestry · 9 gnomAD groups · rare = under 0.1% in that group

Group	People	Missense	Truncating
European (non-Finnish)	~556k	1 in 620	1 in 6,500
South Asian	~43k	1 in 340	1 in 5,300
Remaining (unassigned)	~30k	1 in 310	≈ 1 in 10,000
European (Finnish)	~27k	1 in 680	none seen
Admixed American	~22k	1 in 350	1 in 4,300
African / African American	~21k	1 in 350	1 in 3,900
East Asian	~20k	1 in 430	1 in 2,300
Ashkenazi Jewish	~13k	≈ 1 in 6,000	≈ 1 in 4,400
Middle Eastern	~2,884	1 in 470	≈ 1 in 2,800

Each group is summed over the variants that are rare in that group. Grey ≈ figures rest on fewer than five expected carriers — too few to distinguish groups. Differences mostly reflect how many people were sampled and how much variation each ancestry carries, not disease risk.

Exome coverage (gnomAD): 6 of 7 coding exons well covered in gnomAD exomes; exon 3 marginal. Genomes: all 7.? see exons ↓