For patients with suspected monogenic diabetes diagnosed outside of the neonatal period, we perform exome sequencing followed by targeted analysis of 59 known monogenic diabetes genes (encompassing all genes causing MODY, syndromic forms of diabetes and partial lipodystrophy) and the mitochondrial DNA variant MT-TL1 m.3243A>G.
This technique significantly increases the number of patients with a monogenic diabetes diagnosis compared to targeted sequencing of common cause genes, and identifies patients with syndromic subtypes of monogenic diabetes where there is a low clinical suspicion of the disorder (Colclough et al 2022 Diabetes 71: 530-537, PMID: 34789499 http://link.springer.com/article/10.1007%2Fs00125-013-2962-5).
Which genes/Genetic subtypes are included in the test?
Our monogenic diabetes gene panel includes the genes from the R141 panel (monogenic diabetes) and the familial partial lipodystrophy related genes from the R158 panel (lipodystrophy and severe insulin resistance), details of which can be found on the PanelApp website:
R141: https://panelapp.genomicsengland.co.uk/panels/472/
R158: https://panelapp.genomicsengland.co.uk/panels/546/
A complete list of the genes on our monogenic diabetes panel and the associated phenotypes can be found below.
| Genes | Phenotype |
| GCK | Persistent and mild fasting hyperglycaemia from birth |
| HNF1A and HNF4A | Common forms of MODY that can be treated with sulphonylureas |
| Mitochondrial DNA variant MT-TL1 m.3243A>G | Maternally inherited diabetes and deafness syndrome (MIDD) |
| HNF1B | Renal/urogenital structural abnormalities and diabetes |
| KCNJ11 and ABCC8 | Rare forms of MODY that can be treated with sulphonylureas |
| INS | Non-syndromic autosomal dominant form of MODY |
| NEUROD1, , PDX1, RFX6 | Non-syndromic, autosomal dominant forms of MODY with significantly reduced penetrance for diabetes |
| PCBD1 | Rare, non-syndromic, autosomal recessive form of MODY |
| ZFP57 | Autosomal recessive transient neonatal diabetes and MODY |
| PLAGL1 | Paternal duplication within the 6q24 region causing transient neonatal diabetes and MODY |
| WFS1 and CISD2 | Wolfram syndrome (diabetes, optic atrophy and deafness) |
| AKT2, CAV1, CIDEC, LAMA4, LIPE, LMNA, PCYT1A, PLAAT3, PLIN1, PPARG | Partial lipodystrophy, dyslipidaemia and insulin resistance |
| INSR | Severe insulin resistance without lipodystrophy and dyslipidaemia |
| EPHX1, MFN2, OTULIN, POC5, POLD1, PSMA3, PSMB4, PSMB8, ZMPSTE24 | Lipodystrophy as part of a monogenic syndrome |
| ALMS1 and BLM | Monogenic obesity |
| PCNT | Severe insulin resistance and microcephalic osteodysplastic primordial dwarfism type II (MOPD2) |
| POC1A | Primordial dwarfism syndrome/SOFT syndrome with severe dyslipidaemic insulin resistance |
| WRN | Severe insulin resistance, progeria and cataracts |
| SLC19A2 | Childhood onset diabetes with Thiamine-responsive megaloblastic anemia (TRMA) and hearing loss |
| CEL | Diabetes and exocrine dysfunction |
| PAX6 | Diabetes and aniridia |
| MAFA | Diabetes and insulinomatosis |
| MANF | Diabetes, hypothyroidism, hypogonadism, short stature, ID, obesity, deafness, high myopia, microcephaly and alopecia |
| DUT | Diabetes and bone marrow failure |
| GATA4 and GATA6 | Diabetes and cardiac malformations |
| PPP1R15B, TRMT10A, DUT | Autosomal recessive juvenile-onset diabetes with microcephaly |
| DCAF17 | Woodhouse Sakati syndrome (alopecia, hypogonadism, hearing impairment, diabetes, learning disabilities and extrapyramidal manifestations) |
| ZBTB20 | Primrose syndrome (tall stature, macrocephaly, intellectual disability, disturbed behaviour, unusual facial features, diabetes, deafness, progressive muscle wasting and ectopic calcifications) |
| SMPD4 | Neurodevelopmental disorder with microcephaly, arthrogryposis and structural brain anomalies and diabetes |
| DNAJC3 | Autosomal recessive juvenile-onset diabetes with central and peripheral neurodegeneration |
| FICD | Infancy onset diabetes and severe neurodevelopmental delay |
| DYRK1B | Diabetes and metabolic syndrome |
| PIK3R1 and PRKCE | Diabetes and SHORT syndrome (short stature, hyperextensibility, hernia, ocular depression, Rieger anomaly, and teething delay) |
| SLC29A3 | H syndrome (hyperpigmentation, hypertrichosis, hepatosplenomegaly, heart anomalies, hearing loss, hypogonadism, low height, and hyperglycaemia) and PHID syndrome (pigmented hypertrichosis with insulin dependent diabetes) |
When should I request this test and how much does it cost?
The cost of this test is centrally funded by NHS England for NHSE patients meeting eligibility criteria (see Guidelines for Genetic Testing in MODY // Diabetes Genes). This test for 60 monogenic diabetes genes and m.3243A>G costs £938 for non-NHS referrals. An online probability calculator can predict the likelihood of MODY according to a patient’s clinical characteristics. Please contact the Exeter team for advice regarding the appropriateness of the referral and the best testing strategy for your patient (rduh.betacellgenomics@nhs.net or 01392 408229).
What sample is required?
Venous blood in plastic EDTA bottles (minimum 3ml adults; 2ml children; 1ml neonates) or genomic DNA (minimum of 2µg).
What is the reporting time for this test?
The national target is 42 calendar days.
Can the test identify a pathogenic variant that is known to cause a disease other than diabetes?
Potentially, yes. The test includes the LMNA gene in which pathogenic variants cause a range of phenotypes including partial lipodystrophy, Emery-Dreifuss muscular dystrophy, Charcot-Marie-Tooth disease type 2B, limb girdle muscular dystrophy type 1b, Hutchinson-Gifford Progeria syndrome and dilated cardiomyopathy. More than 200 patients with partial lipodystrophy have been tested in our laboratory and no pathogenic variants causing other phenotypes have been identified. Therefore the likelihood of finding a pathogenic variant that predisposes to, for example, dilated cardiomyopathy is low. For patients with a clinical suspicion of a monogenic form of lipodystrophy, we sequence a panel of genes that includes FBN1. Pathogenic variants in this gene cause Marfan syndrome (MFS) but a subset of variants occurring within the penultimate exon 64 result in a recently recognised syndrome consisting of partial manifestations of Marfan syndrome, a progeroid facial appearance, and clinical features of lipodystrophy (Passarge et al 2016 PMID: 26860060, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4989216/). Analysis of the FBN1 gene will be limited to detecting only known pathogenic variants in exon 64 causing Marfanoid–progeroid–lipodystrophy syndrome to avoid identifying a variant associated with classic MFS.
The test includes genes such as WFS1, HNF1B, GATA6 and the mitochondrial DNA variant m.3243A>G where diabetes occurs as part of a syndromic disorder. The clinical expressivity and penetrance of the extra-pancreatic features are highly variable in these syndromes, and diabetes can be the only presenting feature at the time of testing for monogenic diabetes. This test can therefore potentially result in an unexpected diagnosis of a syndromic disorder (Colclough et al 2022 PMID: 34789499). In these scenarios we will contact the referring clinician to discuss the clinical significance of this result for the patient and their family, and whether referral to a clinical genetics service is required, prior to issuing the formal report.
The test includes the ABCC8 and KCNJ11 genes since pathogenic activating variants cause sulphonylurea responsive diabetes. Congenital hyperinsulinism results from recessive inactivating ABCC8 and KCNJ11 variants; carrier status for hyperinsulinism (identification of a heterozygous inactivating variant) will not be reported.
The test also includes a number of genes that cause recessively inherited forms of monogenic diabetes. It is therefore possible that heterozygous carriers of pathogenic variants in these genes will be identified. These heterozygous variants will not be disease-causing in these individuals and their carrier status will not be reported.
What are the test limitations?
This test will identify the same types of pathogenic variants detected by Sanger sequencing as well as large deletions or duplications previously identified by a separate dosage assay. The test sensitivity is estimated at >97.5% (95% CI). Balanced translocations, inversions and intronic variants located outside of the conserved splice site sequences will not be detected. There are some parts of the human genome that are not amenable to next generation sequencing, including a small part of the GATA6 gene. Pathogenic variants in GATA6 cause pancreatic agenesis with cardiac and other anomalies, but a few patients diagnosed with diabetes in adolescence/adulthood have been reported (De Franco et al 2012 Diabetes 62: 993-997) so this region will be analysed by Sanger sequencing for patients with a cardiac abnormality.
Sequencing a greater number of genes will identify more novel/rare variants whose clinical significance is uncertain. These include variants not seen before or identified in a small number of patients where the causal link to the disease is unproven. The clinical report will only include likely pathogenic variants or variants of uncertain significance where further investigation is recommended (for example by testing diabetic relatives to see if they have the same possible mutation). New clinical or genetic information may change the interpretation of pathogenicity. In accordance with best practice guidelines, known benign variants of no clinical significance are not reported.
Technical information about the test
All protein coding genes are sequenced using the Twist Exome 2.0 plus Comprehensive Exome Spike-in system and the Illumina NovaSeq X Plus next generation sequencing platform.
We sequence the coding regions and intronic regions located within 100 bp upstream and 60 bp downstream of each exon using 150bp paired end reads. We also sequence additional non-coding regions known to harbour pathogenic variants where applicable (i.e. promoters, enhancers and deep intronic sequences). We have developed an additional custom spike-in kit to sequence these regions where they are not covered by the core exome sequencing kit. We also use this custom kit to increase the amount of sequencing performed on regions that are poorly captured by the core exome kit.
We perform targeted analysis of a panel of monogenic diabetes genes using an in-house developed variant calling pipeline using Sentieon software based on GATK best practices, incorporating the Ensembl Variant Effect Predictor (VEP) for annotation and SavvyCNV for detecting copy number variants (CNVs) and large structural variants.
Sequencing overage for all genes within the panel is >98% at 20X. Low coverage of regions will be supplemented by Sanger sequencing for any patients with clinical suspicion of a variant in the low covered region (e.g. cardiac malformation in a patient with low coverage of GATA6 exon 2)
The assay demonstrated 98% sensitivity for SNV detection and 93% for indels across the whole exome, and 99% for SNVs and indels in the targeted genes. An analysis of samples from patients with previously identified pathogenic variants (43 SNVs/indels and 17 CNVs) showed 100% concordance with all variants detected.
All variants of interest are visually inspected using (IGV). Sanger confirmation is only performed for complex insertion/deletion/duplication (indel) variants for the purposes of ensuring that the correct nucleotide change is reported, or where the variant does not meet evidence-based quality parameters. Dosage analysis by MLPA or droplet digital PCR is performed to confirm larger partial and whole gene deletions (excluding the HNF1B whole gene deletion since no false positive deletions have been reported by this assay to date).
