Reading the record for MAGI2 from HGNC, NCBI Gene and Ensembl.Still reading. A first read of a gene can take a while; this page waits up to 115 seconds for it, and its scripts then bring in the page, or a line saying what did not arrive.
The full name, the identifiers, the location and the notes from the sources arrive with the record. Nothing is filled in ahead of it.
Order door
The door to order for MAGI2 opens with the record, which decides which product it carries. The order page itself is open now.
Cytogenetic band 7q21.11NCBI: 7:78,017,055-79,453,667 on the minus strand, GRCh38.p14 (GCF_000001405.40), sequence NC_000007.14, annotation GCF_000001405.40-RS_2025_08 of 2025-08-01Ensembl: 7:78,017,055-79,453,845 on the minus strand, GRCh38.p14 (GCA_000001405.29), release 116Coordinates are one-based with both ends included, as each source reports them.
The protein encoded by this gene interacts with atrophin-1.
NCBI Gene summary
Ready in a moment
Reading NCBI Gene and UniProt.Still reading. A first read of a gene can take a while; this page waits up to 30 seconds for it, and its scripts then bring in the panel, or a line saying what did not arrive.
02 / Transcripts and isoforms
The RNA a design targets
Reading NCBI Datasets and Ensembl.Still reading. A first read of a gene can take a while; this page waits up to 80 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
Reading NCBI Datasets and Ensembl.Still reading. A first read of a gene can take a while; this page waits up to 80 seconds for it, and its scripts then bring in the panel, or a line saying what did not arrive.
03 / Expression by tissue
Where MAGI2 is expressed
Reading GTEx and the Human Protein Atlas.Still reading. A first read of a gene can take a while; this page waits up to 80 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
04 / Protein
The protein MAGI2 encodes
Reading UniProt, InterPro, AlphaFold DB and PDBe.Still reading. A first read of a gene can take a while; this page waits up to 80 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
05 / Interactions
Proteins STRING associates with MAGI2
Reading STRING.Still reading. A first read of a gene can take a while; this page waits up to 50 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
06 / Pathways
Where MAGI2 acts, as Reactome curates it
Reading UniProt and Reactome.Still reading. A first read of a gene can take a while; this page waits up to 60 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
07 / Disease associations
Diseases linked to MAGI2
Reading Open Targets and ClinGen.Still reading. A first read of a gene can take a while; this page waits up to 105 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
08 / Variants
Classified variants of MAGI2
Reading ClinVar.Still reading. A first read of a gene can take a while; this page waits up to 60 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
09 / Constraint
How much variation MAGI2 tolerates
Reading gnomAD and Open Targets.Still reading. A first read of a gene can take a while; this page waits up to 45 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
10 / Orthologs
The same gene in mouse, rat and human
Reading the Alliance, NCBI, Ensembl Compara and RGD.Still reading. A first read of a gene can take a while; this page waits up to 145 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
11 /MicroRNAs
MicroRNAs hosted within MAGI2
Reading Ensembl and miRBase.Still reading. A first read of a gene can take a while; this page waits up to 110 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
12 / Long non-coding RNAs
Long non-coding RNAs at the MAGI2 locus
Reading Ensembl.Still reading. A first read of a gene can take a while; this page waits up to 110 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
13 / Literature
Papers that mention MAGI2
Reading Europe PMC.Still reading. A first read of a gene can take a while; this page waits up to 55 seconds for it, and its scripts then bring in this line, or a line saying what did not arrive.
Ready in a moment
14 / Silencing this gene
From MAGI2 to a sequence that silences it
The AUMsilence™ platform designs the sequences against the human transcripts on this page. Six decisions are yours before it does. What is written under each is AUM's guidance; the timings and the concentrations are in the usage guide below.
01
Choose the region
A knockdown oligonucleotide can sit in the 5' untranslated region, the coding sequence or the 3' untranslated region, and all three are used. The coding sequence and the 3' untranslated region are the usual first choices for an RNase H design; the 5' end near the start codon suits a steric block. The map above shows where each region sits on the isoforms it draws.
02
Cover the isoforms you mean
An exon every isoform carries silences the whole gene; an exon only some isoforms carry silences those and spares the rest. Decide which you want before a sequence is chosen, and check the reference transcript (MANE Select in human; RefSeq Select and Ensembl canonical in mouse and rat) is the one your cells express.
03
Think across species early
The orthologs panel says whether mouse and rat carry the same gene. Whether one oligonucleotide can serve two species is a sequence question, settled at design by matching the candidate against each transcript, not by the protein identity shown there.
04
Check expression in your model
A transcript that is not expressed in your cells cannot show knockdown. Confirm the gene is expressed in the cell type and condition you will use, from your own data or a reference atlas, before the order. The expression panel above gives GTEx's median per tissue for a human gene; for mouse and rat it says that no atlas is on this page yet.
05
Run the controls
A scramble control of the same chemistry, a positive control against a gene known to knock down in your cells, untreated cells, and a mock condition where a transfection reagent is used. Read knockdown at the RNA level first, then at the protein; the usage guide gives the timing and the concentrations to start from.
06
Pick the product
AUMsilence sdASO needs no transfection reagent and works in the cells that resist one. AUMsilence toASO is the transfection-optimised version of the same design, and AUMsiRNA™ is the siRNA route. The selection guide compares them.
For research use only. Not for use in diagnostic or therapeutic procedures.
Gene summary
The protein encoded by this gene interacts with atrophin-1. Atrophin-1 contains a polyglutamine repeat, expansion of which is responsible for dentatorubral and pallidoluysian atrophy. This encoded protein is characterized by two WW domains, a guanylate kinase-like domain, and multiple PDZ domains. It has structural similarity to the membrane-associated guanylate kinase homologue (MAGUK) family.
Provided by RefSeq, Jul 2008, through NCBI Gene. NCBI disclaimer
Protein function
Membrane-associated guanylate kinase, WW and PDZ domain-containing protein 2, Q86UL8
Seems to act as a scaffold molecule at synaptic junctions by assembling neurotransmitter receptors and cell adhesion proteins (By similarity). Plays a role in nerve growth factor (NGF)-induced recruitment of RAPGEF2 to late endosomes and neurite outgrowth (By similarity). May play a role in regulating activin-mediated signaling in neuronal cells (By similarity). Enhances the ability of PTEN to suppress AKT1 activation (PubMed:10760291). Plays a role in receptor-mediated clathrin-dependent endocytosis which is required for ciliogenesis (By similarity)
NCBI Gene summary · NCBI Gene annotation RS_2025_08 · read · NCBI Gene 9863Data from NCBI, provided as is; NCBI's policies and disclaimers apply.
UniProtKB function · 2026_03 · read · UniProt Q86UL8UniProt data are available under the Creative Commons Attribution 4.0 licence.
711 Open Targets disease associations; nephrotic syndrome 15 first, at 0.61.
Open Targets
54 GTEx tissues; the highest median in Artery - Tibial, 6.06 TPM.
GTEx
No annotated microRNA lies within MAGI2 in Ensembl release 116, on GRCh38.p14.
Ensembl · miRBase
8 long non-coding RNA genes overlap MAGI2 in Ensembl release 116, 6 antisense: MAGI2-AS1, MAGI2-AS2, 2 more with a symbol and 4 without.
2,006 PubMed-indexed papers mention MAGI2 at Europe PMC, newest first. Europe PMC ignores letter case, so the count also covers another species' symbol spelled with the same letters.
Europe PMC
1 curated, 0 inferred Reactome pathways for Q86UL8 in human, v97.
Reactome
pLI above 0.9999 and LOEUF 0.348 in gnomAD v4 (GRCh38), on ENST00000354212.9.
gnomAD
The 25 highest-scoring STRING partners at or above a combined score of 0.4, of up to 25 asked for; PTEN, RAPGEF2, CTNNB1 lead.
STRING v12.0
15 RefSeq and 42 Ensembl transcripts on GRCh38.p14; MANE Select NM_012301.4.
NCBI Datasets · Ensembl
Placed on GRCh38.p14 (GCF_000001405.40). MANE Select marks the one transcript RefSeq and Ensembl agree is the reference for this gene.
RefSeq 15 transcripts
coding sequence, tall
untranslated region, thin
non-coding exon
intron, fixed width
Genomic strand: minus. Drawn 5' to 3', so exon 1 sits at the left here and at the highest coordinate on the chromosome.
NM_012301.4NM_012301.4MANE Select
A window on one transcript
One pixel of the map above stands for several bases, and a block too short to see is drawn wider than its own scale, so the map chooses a region and the sequence here chooses the window. Click an exon on a row of the map, or drag across a row; then set the exact start and end below.
These controls are ready in a moment.
No transcript is chosen.
Once a window is chosen this panel shows its length, its G and C count as a percentage of that length, the letters it is made of, the exons it falls in, whether it crosses a junction, and its antisense strand.
Not listed, because the source places them on another assembly only: XM_054359435.1 (not placed on GRCh38.p14); XM_054359436.1 (not placed on GRCh38.p14); XM_054359437.1 (not placed on GRCh38.p14); XM_054359438.1 (not placed on GRCh38.p14); XM_054359439.1 (not placed on GRCh38.p14); XM_054359440.1 (not placed on GRCh38.p14); XM_054359441.1 (not placed on GRCh38.p14); XM_054359442.1 (not placed on GRCh38.p14); XM_054359443.1 (not placed on GRCh38.p14); XM_054359444.1 (not placed on GRCh38.p14); XM_054359445.1 (not placed on GRCh38.p14); XM_054359446.1 (not placed on GRCh38.p14).
NCBI Datasets, RefSeq transcripts · NCBI Datasets 18.38.0; GCF_000001405.40-RS_2025_08 · read · NCBI Gene 9863Data from NCBI, provided as is; NCBI's policies and disclaimers apply.
Ensembl 42 transcripts
A window on one transcript
One pixel of the map above stands for several bases, and a block too short to see is drawn wider than its own scale, so the map chooses a region and the sequence here chooses the window. Click an exon on a row of the map, or drag across a row; then set the exact start and end below.
These controls are ready in a moment.
No transcript is chosen.
Once a window is chosen this panel shows its length, its G and C count as a percentage of that length, the letters it is made of, the exons it falls in, whether it crosses a junction, and its antisense strand.
Lengths are spliced lengths, as each source states them. Exon ranks follow the strand: on a minus-strand gene exon 1 has the highest genomic coordinate. Reference assembly for human: GRCh38.
XM_011516720.4XM_011516720.4
XM_017012847.3XM_017012847.3
XM_017012840.3XM_017012840.3
XM_017012841.3XM_017012841.3
XM_017012842.3XM_017012842.3
XM_017012843.3XM_017012843.3
XM_017012844.3XM_017012844.3
XM_011516718.3XM_011516718.3
XM_017012845.3XM_017012845.3
XM_047421092.1XM_047421092.1
XM_017012846.3XM_017012846.3
XM_047421093.1XM_047421093.1
NM_001301128.2NM_001301128.2
XM_011516728.2XM_011516728.2
scale
Drawn 5' to 3' from each transcript's exons as placed on the reference assembly; exon 1 is the 5' exon on the transcript's own strand. Exon blocks are to scale with each other, except 23 blocks too short to see, widened to a fixed few pixels; introns are drawn at one fixed width whatever their length, so the map is not to scale along the chromosome. Numbers are exon ranks along the strand; a rank is omitted where the exon is too narrow to carry it.
Drawn 5' to 3' from each transcript's exons as placed on the reference assembly; exon 1 is the 5' exon on the transcript's own strand. Exon blocks are to scale with each other, except 46 blocks too short to see, widened to a fixed few pixels; introns are drawn at one fixed width whatever their length, so the map is not to scale along the chromosome. Numbers are exon ranks along the strand; a rank is omitted where the exon is too narrow to carry it.
Mouse Magi2 by 3 of 3 votes; rat Magi2 by 3 of 3 votes. RGD is not answering, so the ortholog list could not be shown. Try again later. Reference e201de37-aa1.
Alliance · NCBI · Ensembl Compara
ClinVar: 760 records for MAGI2, 31 pathogenic or likely pathogenic.
ClinVar
1,455 residues, reviewed Q86UL8; 26 entries from the member databases this page shows along the chain; mean pLDDT 60.5; 5 PDB entries.
