The Translate Pipeline

 

Step 1:

Sequencing of three genes GCK, HNF1A and HNF4A will be performed in the TRANSLATE project based on knowledge that certain variants in these genes result in specific forms of diabetes where precision intervention is possible.

 

Step 2:

The variant rs4149056 located in SLCO1B1 is also included in TRANSLATE. This variant is known to increase risk of myopathy in patients treated with statins (especially Simvastatin) and may therefore have implication for the treatment of hypercholesterolemia with statin.

 

Step 3:

In addition, variants included in the gene risk score for type 1 diabetes are part of TRANSLATE. This risk score can provide information, which can further assist the treating physician in the classification of diabetes type.

 

Step 4:

In the interpretation step of the analysis pipeline, variants will be evaluated in two different loops. One is the automated loop where specific pre-defined variants automatically is linked with a clinical advice based on previous know of the specific variant. Variants with more uncertainty in relation to either a priori knowledge or the quality of the variant will be individually assed by an expert panel.

 

 

Patient above 18 years of age will be offered participation in TRANSLATE and can participate if they sign the informed consent form. Patient inclusion will take place at the following clinics/hospitals:

  • Steno Diabetes Center, Copenhagen
  • Center for Gravide med Diabetes, Obstetrisk Afdeling, Rigshospitalet, Copenhagen
  • Obstetric department, Hvidovre Hospital, Copenhagen
  • Afdeling for Kvindesygdomme, Graviditet og Fødsler, Herlev Hospital, Copenhagen
  • Obstetric department, Nordsjællands Hospital, Copenhagen

 

 

2X10 ml blood will be drawn into two separate EDTA glasses. This step will take place in the Steno Diabetes Center for patients with type 2 diabetes and at the obstetric clinics for women with gestational diabetes. These blood samples will be sent to BGI for further analysis.

 

 

DNA will be extracted. The sample has to include more than 5 ml of whole blood, more than 1ug of DNA, and a DNA concentration higher than 12.5ng/ul. Samples fulfilling these criteria will automatically undergo sequencing.

For samples not fulfilling the above criteria, re-extraction of DNA will be performed.

If sample is damaged to a degree where the above criteria cannot be fulfilled, the treating physician will receive information and can discuss with the patient whether she or he is interested in having new blood sample drawn.

 

 

First collection glass

Purified DNA from one collection glass will undergo whole genome sequencing. The sequencing will be performed on the DNBSEQ-T7 sequencer using paired end (PE) 150. The method is based on the generation of DNA nanoballs and sequencing will be performed to a medium depths of 30X. Only sequencing sample fulfilling the criteria will be transferred to the National Genome Centre (NGC) for further analyses. This transfer will be in the form of fq-files.

 

Second collection glass

The blood sample from the second collection glass will be used for barcoding. Barcoding entails the genotyping of 145 SNPs that will be validated against the sequencing data. Barcoding will be performed at BGI using the method.

 

 

NGC Pipeline

The fq-files will enter the NGC pipeline and will be aligned according to a reference genome [CRCh38] in the form of VCF-files.

Once sequencing files have been assembled, barcoding genotypes will be compared with sequencing results. A concordance rate is above 95% is a strong indication that the barcoding and the sequencing are from the same patient and thus no sample swap has taken place and sequencing data in the form of a VCF-file will be released for Intomics to further analyze the sample.

For samples where there is discordance between barcoding and sequencing data, the treating physician will receive information and can discuss with the patient whether she or he is interested in having new blood sample drawn.

 

 

Annotation will take place at Intomics according to the Ensembl Variant Effect Predictor. Variants located in the investigated genes or variants part of the T1D risk score will be extracted. The nomenclature of the annotated variants is according to the Human Genome Variation Society guidelines (HGVS) (PMID: 26931183). Once annotation is complete, data will be extracted according to the following transcripts:

 

GCK

  • Automated loop: NM_000162.5
  • Expert panel loop: NM_033508.3, NM_001354802.1 og 9

 

HNF1A:

 

HNF4A:

  • Automated loop: NM_175914.4 (PMID: 31815736)
  • Expert panel loop: NM_001287184.2, NM_000457.5, 6 og ENST00000609262.5
The transcripts included in the automated diagnostic loop have been selected in order to annotate immediately actionable variants in accordance with the most commonly used annotation for monogenic diabetes. However, as these transcript to do not contain all regions of the gene for all tissues.
Therefore, to ensure that variants classified as coding according to a transcript different from the one selected for the automated loop, we will annotate variants all possible coding variants. Variants only found to be coding in transcripts not applied in the automated loop, will be assessed individually by the expert panel who will provide the treating physician with a clinically relevant interpretation of the identified variant.
 

Sequencing Quality

Sequencing quality criteria for variants exists at two levels. For variants entering the automatic diagnostic loop, a set of strict criteria has been defined and only variants fulfilling the following criteria will be further considered for the automated loop. The criteria for such variants are:
  • Depths of ≥20x
  • Q30 for the site
  • Allelic ratio for heterozygotic variants between 0.35 and 0.65
  • Allelic ratio for homozygotic variants above 0.9
Variants not fulfilling these criteria will (if they are included among immediately or potentially actionable variants, see below) be submitted to the expert panel, who will inspect the variants individually and based on the quality and the location of the variant, provide the treating physician with a clinically applicable interpretation of the variant.

 

 

The filtered genotypes are returned in a report.

Types of variants

All type of variants encountered will been classified as either:

  • Immediately actionable
  • Potentially actionable
  • Non-actionable variants

Immediately actionable variants fulfilling the quality criteria will automatically provide the treating physician with a clinically applicable interpretation.

Potentially actionable variants or immediately actionable variants not fulfilling the quality criteria, will be evaluated by the expert panel, who will provide a clinically useful interpretation to the treating physician.

Non-actionable variants will automatically provide the treating physician with the information that genetics should not be considered in the treatment of the patient.

 

Variants classified as immediately actionable

Variants has been classified as immediately actionable based on:

  • Variants known to cause loss of function for the particular gene (stop mutations, frame-shift, exon deletion)
  • Variants previously shown to co-segregate with MODY in families
  • Variants classified as pathogenic according to either Clinvar, ACMG or HGMD
  • Variants not identified in non-diabetic individuals
  • Variants with in vitro functional impact

Criteria for selecting potentially actionable variants

Variants has been classified as potentially actionable based on:

  • Variants identified in either MODY patients or patients with type 2 diabetes
  • Variants less frequent in non-diabetic individuals compared to type 2 diabetes
  • Variant with suggestive in vitro functional impact
  • Variants classified as pathogenic or VUS in Clinvar, ACMG or HGMD
  • Variants being novel

 Criteria for selecting non-actionable variants

  • Variants has been classified as non-actionable based on:
  • Variants identified in several non-diabetic individuals
  • Variants with a frequency too high to have a strong functional impact
  • Variants not affecting the amino acid sequence of the gene or not located in known important regulatory regions
  • Variants classified as benign/non-disease causing in Clinvar, ACMG or HGMD)

The specific description of the selected non-actionable variants can be found in Appendix 2

If patients carry more than one immediately actionable variant or carry one immediately as well as a potentially actionable variant, the interpretation and clinical recommendation will be performed by the expert panel.