Early-infantile DEE: genes and variants

Genes linked to Early-infantile DEE

• SCN1A: Sodium channel protein type 1 subunit alpha

  Its sodium current is especially important for reliable firing of inhibitory interneurons and therefore for balancing excitation across neural networks. Loss-of-function variants are the major cause of Dravet syndrome, while other variants cause GEFS+ or familial hemiplegic migraine.

  422 disease-causing and 758 uncertain variants in SCN1A are linked to Early-infantile DEE.

• KCNQ2: Potassium voltage-gated channel subfamily KQT member 2

  Together with KCNQ3, its slowly activating potassium current provides much of the neuronal M-current that suppresses repetitive firing. Pathogenic variants cause a spectrum from self-limited familial neonatal epilepsy to severe developmental and epileptic encephalopathy.

  199 disease-causing and 446 uncertain variants in KCNQ2 are linked to Early-infantile DEE.

• SCN8A: Sodium channel protein type 8 subunit alpha

  The protein forms Nav1.6, a voltage-gated sodium channel that sets the threshold and propagation of neuronal action potentials. It is widely important for neuronal excitability, and SCN8A variants are associated with developmental and epileptic encephalopathies.

  90 disease-causing and 648 uncertain variants in SCN8A are linked to Early-infantile DEE.

• STXBP1: Syntaxin-binding protein 1

  It controls SNARE-complex assembly and synaptic-vesicle fusion, making it essential for rapid neurotransmitter release. Haploinsufficiency causes STXBP1-related neurodevelopmental disorder with developmental impairment, epilepsy, movement abnormalities, and intellectual disability.

  45 disease-causing and 161 uncertain variants in STXBP1 are linked to Early-infantile DEE.

• HCN1: Potassium/sodium hyperpolarization-activated cyclic nucleotide-gated channel 1

  The protein forms a hyperpolarization-activated channel that conducts both potassium and sodium ions. It contributes to pacemaker currents and the neuronal I(h) current that shapes excitability, and HCN1 variants are associated with developmental epilepsy syndromes.

  14 disease-causing and 285 uncertain variants in HCN1 are linked to Early-infantile DEE.

• CACNA2D2: Voltage-dependent calcium channel subunit alpha-2/delta-2

  It promotes trafficking and functional expression of voltage-gated calcium-channel complexes and is particularly important in cerebellar neurons. Biallelic loss-of-function variants can cause developmental epileptic encephalopathy with cerebellar atrophy and ataxia.

  0 disease-causing and 35 uncertain variants in CACNA2D2 are linked to Early-infantile DEE.

Where Early-infantile DEE variants cluster

• KCNQ2 Mediates interaction with SLC5A3/SMIT1 (positions 222–323): 85 of 199 disease-causing changes, 3.6× more than its size predicts.
• KCNQ2 Segment S4 (positions 197–215): 24 of 199 disease-causing changes, 5.5× more than its size predicts.
• SCN1A IV (positions 1523–1821): 107 of 422 disease-causing changes, 1.7× more than its size predicts.
• HCN1 Segment S6 (positions 372–392): 6 of 14 disease-causing changes, 18.2× more than its size predicts.
• SCN1A S6 of repeat II (positions 966–992): 19 of 422 disease-causing changes, 3.4× more than its size predicts.

Known disease-causing variants in Early-infantile DEE

Showing 60 of 770.

Uncertain variants in Early-infantile DEE that look disease-causing

Which prediction tools work for Early-infantile DEE

How often each tool ranks a disease-causing variant above a harmless one (AUROC × 100).

• REVEL: 97 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• AlphaGenome (regulatory): 94 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• CATVariant: 92 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• MetaLR: 91 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• AlphaMissense: 88 out of 100
• PolyPhen-2: 88 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• SIFT: 87 out of 100
• ESM1b (LLR): 86 out of 100
• CADD: 85 out of 100
• EVE: 83 out of 100
• MutPred2: 81 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• AlphaGenome (splicing): 67 out of 100 (learned from overlapping clinical labels, so this is optimistic)
• phyloP: 61 out of 100

Same protein, different disease

• Severe myoclonic epilepsy in infancy is also caused by SCN1A variants; they fall in the same places as the Early-infantile DEE variants (213 disease-causing).
• Generalized epilepsy with febrile seizures plus is also caused by SCN1A variants; they fall in the same places as the Early-infantile DEE variants (76 disease-causing).
• Migraine, familial hemiplegic, 1 is also caused by SCN1A variants; they fall in the same places as the Early-infantile DEE variants (20 disease-causing).
• Seizures, benign familial neonatal, 1 is also caused by KCNQ2 variants; they fall in the same places as the Early-infantile DEE variants (57 disease-causing).
• Cognitive impairment with or without cerebellar ataxia is also caused by SCN8A variants; they fall partly in the same places as the Early-infantile DEE variants (22 disease-causing).
• Seizures, benign familial infantile, 3 is also caused by SCN8A variants; they fall mostly in different places as the Early-infantile DEE variants (12 disease-causing).
• Complex neurodevelopmental disorder is also caused by SCN8A variants; they fall partly in the same places as the Early-infantile DEE variants (8 disease-causing).
• Autosomal recessive inheritance is also caused by SCN8A variants; they fall mostly in different places as the Early-infantile DEE variants (3 disease-causing).
• Myoclonus, familial, 2 is also caused by SCN8A variants; they fall partly in the same places as the Early-infantile DEE variants (3 disease-causing).
• Generalized epilepsy with febrile seizures plus is also caused by HCN1 variants; they fall mostly in different places as the Early-infantile DEE variants (10 disease-causing).

Diseases related to Early-infantile DEE

• Genetic developmental and epileptic encephalopathy, also linked to CACNA2D2, HCN1, KCNQ2, SCN1A and 2 more
• Epilepsy, also linked to CACNA2D2, KCNQ2, SCN1A, SCN8A and 1 more
• Severe myoclonic epilepsy in infancy, also linked to HCN1 and SCN1A
• Amyotrophic lateral sclerosis, also linked to SCN1A and SCN8A
• Generalized epilepsy with febrile seizures plus, also linked to HCN1 and SCN1A
• Cardiac arrhythmia, also linked to SCN1A and SCN8A
• Autosomal dominant epilepsy, also linked to KCNQ2 and SCN1A
• Lennox-Gastaut syndrome, also linked to SCN1A and SCN8A
• Undetermined early-onset epileptic encephalopathy, also linked to HCN1 and SCN8A
• Seizures, benign familial infantile, 3, also linked to SCN8A
• Migraine, familial hemiplegic, 1, also linked to SCN1A
• Seizures, benign familial neonatal, 1, also linked to KCNQ2

Frequently asked questions

Which genes are linked to Early-infantile DEE?

In CATVariant, Early-infantile DEE is linked to 6 analyzed proteins: SCN1A (Sodium channel protein type 1 subunit alpha), KCNQ2 (Potassium voltage-gated channel subfamily KQT member 2), SCN8A (Sodium channel protein type 8 subunit alpha), STXBP1 (Syntaxin-binding protein 1), HCN1 (Potassium/sodium hyperpolarization-activated cyclic nucleotide-gated channel 1) and CACNA2D2 (Voltage-dependent calcium channel subunit alpha-2/delta-2).

How many genetic variants are linked to Early-infantile DEE?

3,336 variants: 770 are classified as disease-causing (pathogenic or likely pathogenic) in ClinVar and 2,333 are of uncertain significance or have conflicting reports.

Which uncertain variants in Early-infantile DEE look disease-causing?

6 uncertain variants reach the likely-pathogenic range of the ACMG/AMP points scale on computable evidence, for example SCN8A R1626H, HCN1 A387S, STXBP1 R190Q, SCN8A R1872Q and KCNQ2 L307R. These are leads for expert review, not diagnoses.

Which variant effect predictor works best for Early-infantile DEE?

Among tools not trained on clinical labels, AlphaMissense separates this disease's known disease-causing variants from harmless ones best (AUROC 0.88, based on 627 disease-causing and 84 harmless variants).

About this data

Variant–disease links come from ClinVar, Open Targets and UniProt, pooled from the latest public CATVariant analysis of each human protein. Evidence scores use the ACMG/AMP Bayesian points scale with computable criteria only (position among known disease variants, rarity in gnomAD, calibrated predictors, deep mutational scanning); there is no family or patient data, so they prioritise variants for expert review and never classify them.

Download every variant as CSV · Browse all diseases · Methods · About the Center