Notes from NGS applications – Stephen Kingsmore, Fowzan Alkuraya, Hakon Hakonarson

I’m dumping my notes as I’ve done for other conferences – obvious mistakes are obviously my fault, and not those of the speakers.

Fowzan Alkuraya, Alfaisal University

Case 1 – 13 month old girl – developmental delay.
Unrelated parents
MRI brain atrophy
Karyotype: 45,X (non mosaic turner syndrome)

In past, would have said “atypical Turner Syndrome”
Now, that’s not good enough – we can find something else. “Atypical” should really not be used anymore – there’s probably more than one “lesion”

Exome sequencing:
ADRA2B – Arg222* -> homozygous truncating mutation.

Lesson:¬†Don’t assume – there’s no excuse for “atypical” in the genomics era.

Case 2 – 4 year old suspected autism.
Un-contributory family history with healthy brother – should raise flag: autism is more common in boys than girls. Mendelian form of Autism?
Documented cognitive impairment – otherwise normal.
All guidelines : molecular karyotype: de novo 300kb deletion on chr10.
Is it pathogenic?
Use Decipher database of structural variants. ***** why don’t we use this?
found a match.

Were conducting a study that included clinical genomics approach, and exome sequencing found:
homozygous mutation in CC2D1A. Skipping exon 6. Not in exac, but found in saudi Arabians. (1 in 500) Known to be entirely correlated with mental development.

Beware of founder mutations in different ethnic groups.
Exome sequencing in parallel with molecular karyotyping for neurodevelopmental disorders.

But, when do we stop? Do we always need Exome sequencing?

Case 3 – (consanguineous) Couple lost two children with severe lactic acidosis (Severe, unexplained)
First child died on 2nd day
Second child died within hours.

normal electron transport chain. Sequencing of candidate genes was negative. clinical Exome sequencing: negative.

clinical Whole Genome sequencing: Negative.

Research grade exome sequencing: Found a splicing mutation in ECHS1, a known source of acidosis

Severe reduction in NMD.

30-50% of cases in exome sequencing remain without diagnosis. Are we normally missing the mutation at the capture and sequencing stage, or at the intepretaiotn stage?

Analyzed 33 cases with negative clinical exome/genome sequencing. Found it in 29 cases.

In 18 cases, gene was novel or within 6 months at time of diagnostics.
probably not reported
In 11 cases were in known genes.

Clinical directors are probably not reported because of filtering and interpretation issues.

If you have a novel mutation, it’s likely to be missed by clinical sequencing.

Stephen Kingsmore, Rady Children’s Hospital:

2 cases:
first, ad birth, acute liver failure,
surgically corrected
spine defects, renal defects, surgically correctable.
Doing well until day 40, when he started to develop liver disfunction. Diagnostic workup was unrevealing.
On day 55, Rady would brought in. Race against the clock.
Whole Genome Sequencing time cut to just 26 hours.

1. consent at time 0:00
sample transport
Dna isolation 1 hour
18 hour genome sequencing, completed at time 24:30

40x genome. 120,000,00,000 bases
2,8M bases
5,1M variants
1.3M variants1% filter applied.
1,3k pathogenic or likely pathogenic.
2 variants that could cause the 341 conditions (below), both in the same gene: perforin 1.

– very typical, but has to be done fast. FPGA informatincs.

ACMG guidelines on how build cases.

Focus on pathogenic and likely pathogenic

Big issue: what are the issues that are related to the genotypd of intrerest. Used Phenomizer, etc, were able to narrow down to 341 conditions that may match the symptions.

1st variant was vary rare.
2nd variant was in 3%.

If second variant is in trans with another pathogenic, it’s likely pathogenic as well.
Provisional Diagnosis – FDA gave permission to give a verbal putative diagnosis under cases where a child’s life is in imininent danger.

Confirmatory testing was done, and the diagnosis was positive. Fortunately, there is a treatment, and the child is now thriving. Does still have the disorder,which may require a bone transplant, but the child’s life was saved.

Case 2: firstborn with transient hypoglycemia. Transient to Nicu.

At 1 month, the nurse practitioner noticed low blood suger – hyperinsulinemia.

Similar numbers to previous, whittled down to 160 conditions, with only 1 variant that matched that disease.

Known pathogenic mutation : ABCC8.

Recessive condition, inherited from father. There is also: Focal hyperinsulinism, which presents from father (uniparental disomy)

Second event was a de novo mutation in the child, which was shown to be only at the head of the pancreas – so were able to remove the damaged segment of the pancreas.

Pancreatectomy was scheduled.

TOtal time: 7 days from start to cure.

Avoided major morbidity – probably major neurologic damaged.

Does it scale?

35 cases: cohort. 57% rate percent diagnosis. In contrast, 9% are cured by standard tests.

These cases were cherry picked as having likely genetic diseases., but still demonstrates power.

2nd version:
80 cases cohort : 58% diagnosis ate.

Brand new info: Kansas city test. However, trial was discontinued because it was obvious that diagnosis was working: 15% rate in normal tests. 41% in clinical exome based tests.

Makes a significant impact in all aspects of care.

For every child tested, 2,9 quality years improvement. or, $3500 per quality years.

Hakon Hakonarson – Children’s Hospital¬†of Pensylvania

Centre for Applied Genomics at CHOP. Collaborate with Penn.

Case from Lipid Cohort. Familial form of lipid disease. 1700 subjects, 900 families.

Case: 55 year old man – phenotype described (no fat, mild diabetes, lipoprotein panel appeared normal.) [Missing much of it – don’t know the terms]

Many features overlapped with adult progeria.

Initial genetic analysis: turned out to be homozygous for PLIN1 and heterozygous for WRN.
Balanced translocation t(8;10) as well.

Pedigree shown. Two brothers, both with much more gentle phenotypes, both had liver issues.

Goal became to map the breakpoint: were there any additional genes or elements contributing to the phenotype? The condition is far more advanced in proband.

Used Linked Read technology for translocation breakpoint mapping. (Quick review of barcoding for this technology) Gel beads in emulsion.

Fine mapping of region: near cyp26C1 and CYP26A1 and ADHFE1 on the other side.

No single gene jumped out.

Many hypotheses were considered – not clear what was going on. Next step is investigation into WRN. Cell lines used for WRN activity, protein expression and transcript assessment.

Asses changes in genes near breakpoint.

Not totally solved, but very interesting case.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.