Translating Rapid Whole Genome Sequences into Precision Medicine for - - PowerPoint PPT Presentation

Translating Rapid Whole Genome Sequences into Precision Medicine for Infants in Intensive Care Units

Stephen F. Kingsmore, MB, ChB, DSc, FRCPath, President, Rady Children’s Institute for Genomic Medicine, San Diego

"Tonight, I'm launching a new Precision Medicine Initiative....to give all

• f us access to the personalized information we need to keep
• urselves and our families healthier….a new era of medicine….that

delivers the right treatment at the right time."

National Adoption of Genomic Medicine

14% of US newborns admitted to a NICU

Rationale

Economics Work: Cost of care $4000 per day Leading cause

• f

Death in NICU & PICU Conventional testing Too slow To guide care

Diagnosis of 8250 Genetic Diseases in NICUs

Conventional Testing Too Slow For Optimal NICU Outcomes

Proportion Surviving Days of Life

Why single gene diseases?

• Simple, deterministic genetics:

– 1 or 2 pathogenic variants in/near 1 gene that causes disease with symptoms similar to that in the infant = necessary and sufficient

• 50yr medical genetic infrastructure

– Medical geneticists, genetic counselors, public health services

• Orphan drug companies, gene therapies

Baby CMH487: acute liver failure

00:00

Gap 2: Making a differential diagnosis: Complexity of Genetic Diseases in Symptomatic Infants

Typical presentations of 4,645 known genetic diseases Partial presentations of 4,645 known genetic diseases Atypical presentations of 4,645 known genetic diseases 20+ novel genetic diseases discovered per month Genetic disease mimicking a non-genetic disease Genetic disease complicating a non-genetic disease Two concomitant genetic diseases 3,643 named, uncloned genetic diseases Stereotyped presentations due to partially developed organ systems and homeostatic responses

April 10, 2013

Baby CMH487

00:00

Hydrocele testis Infra-orbital crease Maternal diabetes Low-set, posteriorly rotated ears Feeding difficulties in infancy Ventilator dependence Single umbilical artery Cholestasis Thrombocytopenia Prolonged partial thromboplastin time Prolonged prothrombin time Chronic lung disease Hypertelorism Thoracolumbar scoliosis Bronchodysplasia Omphalocele Chin dimple Duplicated collecting system Ventricular hypertrophy Nevus flammeus Gastroesophageal reflux

12,000 HPO terms x 5,500 genetic diseases x 3,500 disease genes

April 10, 2013

Baby CMH487

00:00

HP:0009800 Maternal diabetes HP:0001539 Omphalocele HP:0008872 Feeding difficulties in infancy HP:0006533 Bronchodysplasia HP:0005946 Ventilator dependence HP:0006528 Chronic lung disease HP:0002020 Gastroesophageal reflux HP:0002944 Thoracolumbar scoliosis HP:0000081 Duplicated collecting system HP:0000034 Hydrocele testis HP:0001195 Single umbilical artery HP:0100876 Infra-orbital crease HP:0000368 Low-set, posteriorly rotated ears HP:0000316 Hypertelorism HP:0010751 Chin dimple HP:0001052 Nevus flammeus HP:0001396 Cholestasis HP:0001873 Thrombocytopenia HP:0003645 Prolonged partial thromboplastin time HP:0008151 Prolonged prothrombin time HP:0001714 Ventricular hypertrophy

Parents gave consent

00:00

Blood sample from mum, dad and baby

00:00

San Diego Synergy: Illumina + Edico + Rady Children’s

Transport to Institute

00:02

Robot isolates genomic DNA

01:00

Robot prepares DNA for sequencing

06:00

18 hour genome sequencing

24:30

Each of my 37 trillion cells contains 2 genomes of 3.2 billion DNA letters

We are fearfully and wonderfully made. Psalm 139

Scope: Comparison of Short Read Genome and Exome Sequencing

Pros of Genomes:

• One day faster
• Samples 90% of genome
• Reasonably good deletion

structural variant detection

Pros of Exomes:

• Cost ¼ that of genomes

Disadvantages of both:

• Limited phasing; requires trios
• Don’t detect insertion structural

variation or repeat expansions Gene Whole Genome Sequence Whole Exome Sequence = 2%

Co Cost an and T Time-to-Res esul ult o

• f Genomic S

Sequenc uencing ng O Opt ptions ns

48+ 48+ 40X 4000 Trio Genome 48+ Hr 2x80nt

Gap 4:

24:30

Infant CMH487

9 TB/day 6GB/sec 3 TB/day 600 Mb/sec 5TB/day 1 GB/sec

Remote backup Genome sequencers 24 x 7, 365 days/yr High performance NAS Compute cluster In-house researchers Single monitoring dashboard 75 TB

10 Gbps 10 Gbps 10 Gbps Simultaneous writes

500TB

100 cores 1TB/day External Collaborators

500TB Public cloud

AWS Google Azure 25 yr archival 1 Gbps http

Single global name space

Replication 2 yrs worth storage

Local object storage

1 Gbps http

Compute & Storage

Variant database

GPFS

24:45

Infant CMH487

25:00

Infant CMH487

Improved Analytic performance of WGS (nucleotide variants; not structural variants)

Sample Yeild (GB) Site Pipeline A DRAGEN GSNAP/GATK-1.6/noVQSR DRAGEN GSNAP/GATK-3.2/noVQSR Essex CMH NA12878 NA12878 65

a

143 Analytic Sensitivity Analytic Specificity 99.93% 99.87% 99.54% 98.57% 99.42% 99.46% 97.29% 95.35% 45X 20X Coverage

25:01

Infant CMH487

Hemoglobin-β DNA code

Normal red blood cell

Does this DNA letter change cause a genetic disease?

Sickle cell

Variant pathogenicity categories

Genet Med. 2015 Mar 5. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of ACMG and AMP. Richards S, et al.

• Same amino acid change as previously established pathogenic variant
• De novo in a patient with the disease and no family history
• Functional studies show damaging effect on the gene
• Prevalence in affected individuals significantly greater than controls

• Located in mutational hot spot/functional domain without benign variation
• Extremely low frequency in Exome Sequencing or 1000 Genomes Projects
• For recessive disorders, detected in trans with a pathogenic variant
• Protein length changed by in-frame indel in nonrepeat region or stop-loss
• Novel missense at amino acid where different missense known to be pathogenic
• Assumed de novo, but without confirmation of paternity and maternity

• Missense variant in gene with low rate of benign missense variants and where

• Multiple computational tools call deleterious
• Phenotype highly specific for disease with single genetic etiology
• Reputable source reports as pathogenic, but unpublished

Category CRITERIA Pathogenic 1VS + (1S or 2M/Supp) 2 Strong 1S + 3M or (2M+2Supp) Likely Pathogenic 1 VS/S + 1 M 1 S + (1 M or 2 Supp) 3 M 2 M + 2 Supp 1 M + 4 Supp

25:41

Infant CMH487

341 Possible Diagnoses

25:42

Computer-Aided Comprehensive Differential Diagnosis

Infant CMH487 Diagnosis

25:43

• Perforin 1 heterozygous c.272C>T [p.Ala91Val] P,

supported by functional studies

• Perforin 1 heterozygous c.1310C>T [p.Ala437Val] LP,

supported by case-control studies

• Diagnosis: Hemophagocytic lymphohistiocytosis type 2

26:00

• Confirmatory testing
• Precision Medicine:

– Meds. D/Cd – IV Ig and steroids

Hemophagocytic Lymphohistiocytosis Diagnostic Criteria (need 5) Present in CMH487 Fever No Hepatomegaly or splenomegaly Modest (1) 2 cytopenias: hemoglobin<9 g/dL, platelets <100,000/mm3, ANC <1000 Yes (2) Serum ferritin >500 ng/mL Yes (3) Serum triglyceride >265 mg/dL or fibrinogen <150 mg/mL Yes (4) Absent/decreased natural killer cell assay Yes, after Dx (5) Soluble IL2 receptor (CD25) >2,400 units/mL Not done Hemophagocytosis without malignancy Not done

Outcome

• Coagulopathy resolved on d. 7
• 7 surgeries for correction of congenital anomalies
• He is now 32 months old, normal liver function
• 72 quality adjusted life years saved

Diagnostic Utility: Meta-Analysis of 9 Studies of Short Read Exome & Genome Sequencing in Children

• 1. Overall Diagnosis Rate = 28% = most effective

method for making molecular diagnosis of childhood genetic disease

Chromosomal microarray (current 1st tier test for genetic disease Dx): diagnostic rate ~15%

• 3. de novo Mutations: The most common

mechanism of genetic disease diagnosis

Willig LK, et al. Science Trans. Med. April 2015

35 NICU / PICU infants with

likely genetic disease (Kansas City)

57% (20)

By rapid WGS

9% (3)

By standard methods

Molecular Diagnosis

Stark Z, et al. Genetics in Medicine 3/3/2016

80 infants under care for likely

genetic disease (Melbourne)

14% (11)

By standard methods

58% (46)

By Exome Seq

Molecular Diagnosis

Explore use

• f

genomic information for broadening understanding of diseases identified in the newborn period

Newborn Sequencing In Genomic medicine and public HealTh

65 Randomized

NICU infants of age < 4 months with clinical features suggesting a genetic disease 33 Infants with Standard Tests 32 Standard Tests and Trio Rapid WGS

NSIGHT Randomized Controlled Study of rapid WGS in NICU infants

5 Cross-overs

Molecular Diagnosis 5 (15%) 13 (41%) p<0.05

Study terminated early due to loss

• f equipoise of neonatologists

Time to Diagnosis 65 days 16 days

2.9 QALYs per family tested

YES: 2.9 QALYs per newborn genome Cost per QALY: $3500

Kansas City Melbourne Change in care 13 (37%) 16 (20%) Life-saving treatment 1 (3%) 2 (3%) Major morbidity avoided 3 (9%) 1 (1%) Major Procedure Change 3 (9%) 4 (5%) Palliative Care Guidance 6 (17%) 0 (0%) Medication Change 4 (11%) 8 (10%) NICU stay decreased by >1 month 1 (3%) 0 (0%) Parent or sibling diagnosed 1 (3%) 10 (13%) Diet Change 2 (6%) 2 (2%) Complication monitoring 1 (3%) 11 (14%)

CMH5033, a 3-week-old male with isolated, symptomatic atrial flutter, and his parents

CMH5033: 3 week old male with atrial tachydysrhythmias

DOL2: Local NICU; Atrial tachycardia; Premature atrial contractions; Some non- conducted P waves; Heart rate normal; Discharged DOL7. 2 weeks old: Outpatient; tachypneic, gained 1 oz. EKG: Atrial tach 190-230. Some non- conducted P waves. Admitted CMH NICU. Episodes atrial flutter 217 bpm with pallor, thready pulse. Cardioverted x 2. ECG: 6/2: Vent.Rate:164BPM, Atrial Rate:178BPM, Normal P-R Interval, QRS Interval and QTc. Rx: Amiodarone, esmolol, flecainide. ECHO: Structurally normal. Patent foramen ovale. Ectopic beats. During sinus beats, LV ejection fraction 52%. During ectopic beats, abnormal septal motion. Amiodarone and esmolol D/Cd after flecainide therapeutic. 3 weeks old: Enrolled for WGS; Hemodynamically stable; Eating well; Discharged.

Trio Whole Genome Sequencing

• Day 1: Enrolled
• Day 2: Blood samples from trio
• Day 4: Finished 2 x 100 nt, HiSeq 2500, proband, 108GB &

mother 109GB

• Day 4: Proband VCF analyzed
• Day 5: Finished father WGS (89GB), re-analysis of trio
• Day 9: Laboratory director completed analysis/interpretation

Automated Differential Diagnosis

• HPO terms: atrial flutter (mandatory), AFib (present),

multifocal atrial tachycardia (present), 1 gene (SCN5A)

• Phenomizer: terms present: P<0.05; 259 genes
• 17 OMIM AFib loci
• These genes harbor 1 possibly pathogenic variant in the

proband with minor allele frequency<0.5%

IGV: Proband, 3:38591853-38591853 A > G

Good coverage, bidirectional reads call variant, clean alignments

Filters: Potentially pathogenic variant(s) with allele frequency <0.5% in genes assoc. with atrial fibrillation: 1 heterozygous variant

Inherited from father (As far as we know he is unaffected)

Proband Mother Father

Complication: There are 9 OMIM SCN5A Phenotypes

• 1. Non-synonymous

hgvs_c: NM_000335.4:c.6007T>C, hgvs_p: NP_000326.2:p.Phe2003Leu hgvs_c: NM_001099404.1:c.6010T>C, hgvs_p: NP_001092874.1:p.Phe2004Leu hgvs_c: NM_001099405.1:c.5956T>C, hgvs_p: NP_001092875.1:p.Phe1986Leu hgvs_c: NM_001160160.1:c.5911T>C, hgvs_p: NP_001153632.1:p.Phe1971Leu hgvs_c: NM_001160161.1:c.5848T>C, hgvs_p: NP_001153633.1:p.Phe1950Leu hgvs_c: NM_198056.2:c.6010T>C, hgvs_p: NP_932173.1:p.Phe2004Leu In silico tools largely predict to be benign SIFT: tolerated (0.65) MutPred: Medium risk Grantham score: 22 (range 5 – 215) LRT: Neutral FATHMM: Damaging (-3.6)

• 2. Occurs in a functional domain of the protein:

Unstructured COOH terminus is involved in inactivation gating 5 variants in this region have physiologic phenotypes

• 3. Occurs in a 20 AA region that is well conserved;

Albeit PolyPhen2 = benign (0.001), PhyloP 0.04 And, Dog and Chinchilla have leucine at this position

Support for Pathogenicity of 3:38,591,853-38,591,853A>G

• 4. Allele frequency – rare, but not extremely low

CMH MAF 0.24%, 11 het., 0 hom., / 4666 total alleles EVS MAF European American 0.31% ExAC MAF 0.20% dbSNP: rs41311117 MAF 0.14%

• 5. Reported in patients with sudden infant death syndrome, sudden cardiac death, Brugada syndrome,

and in apparently healthy individuals…Oleson et al. Circ Cardiovasc Genet 2012 5:450-459 – F2004L father and proband had early onset AF

• 6. Altered electrophysiologic effects in 2 manuscripts

Wang: “F2004L alone is not sufficient to evoke arrhythmia susceptibility, and other factors are probably needed for full pathophysiological expression” Bebarova: “We believe that the F2004L loss-of-function variant is a disease-associated mutant”. “…the question arises why the pathogenic potential of this SCN5A variant is so variable. Undoubtedly, modifier genes could play a role.”

Support for Pathogenicity of 3:38,591,853-38,591,853A>G

• Circulation. 2007;115:368-376

Genet Med. 2015 Mar 5. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of ACMG and AMP. Richards S, et al.

• Same amino acid change as previously established pathogenic variant
• De novo in a patient with the disease and no family history
• Functional studies show damaging effect on the gene
• Prevalence in affected individuals significantly greater than controls

• Located in mutational hot spot/functional domain without benign variation
• Extremely low frequency in Exome Sequencing or 1000 Genomes Projects
• For recessive disorders, detected in trans with a pathogenic variant
• Protein length changed by in-frame indel in nonrepeat region or stop-loss
• Novel missense at amino acid where different missense known to be pathogenic
• Assumed de novo, but without confirmation of paternity and maternity

• Cosegregation with disease in multiple affected family members in gene known to cause disease
• Missense variant in gene with low rate of benign missense variants and where missense variants commonly

• Multiple computational tools call deleterious
• Phenotype highly specific for disease with single genetic etiology
• Reputable source reports as pathogenic, but unpublished

Interpretation 1: Likely benign (1 strong + 2 supporting)

• BS1: Allele frequency greater than expected
• BP6: LMM assertion is likely benign in Clinvar
• BP4: In silico predictions

Interpretation 2: Pathogenic (2 strong)

• Functional studies show damaging effect on the gene
• Prevalence in affected individuals significantly greater than controls

Questions to be answered at the end

• f the presentation:

Based on the aggregate evidence, do you believe SCN5A p.Phe2004Leu is: Pathogenic Likely Pathogenic Likely Benign Benign 1 2% 22 61% 13 36% 0 0%

Questions to be answered at the end

• f the presentation:

Based on the aggregate evidence, do you believe that, in this case, SCN5A p.Phe2004Leu is: Causative Likely causative A risk factor Unrelated 1 3% 10 37% 15 55% 1 3%

Questions to be answered at the end

• f the presentation:

Based on these two answers, in this research study, would you report SCN5A p.Phe2004Leu as: Diagnostic No report, since non-diagnostic Report, clinical correlation warranted 1 2% 8 19% 32 78%

Summary

• 1. 15% of the 15% of babies admitted to ICUs may benefit from rapid genome

sequencing

• 2. Case studies show WGS saved 2.9 QALYs per test in ICU/ICU infants
• 3. Clinical utility is multi-dimensional and requires novel multidisciplinary teams e.g.

for precision palliative care

• 4. Additional clinical trials are needed to define optimal methods and delineate

infants who will benefit

• 5. Systematic healthcare changes needed for broad adoption of genomic medicine in

ICU infants

Proposal for the SADS Foundation

• (Blame Jim Perry)
• A randomized, controlled study of rapid WGS

as a first tier test in Code Blue children

Support: Children’s Mercy Hospital Rady Children’s Hosptial NICHD NHGRI NIDDK

The LORD is my creativity, Psalm 27

Josh Petrikin Laurel Willig Carol Saunders John Lantos Neil Miller Emily Farrow

Lance Prince MD PhD Gail Knight MD Farhad Imam MD PhD Nathaly Sweeney MD Nicole Coufman MD PhD Mallory Boutin Julie Ryu MD Cynthia Kuelbs MD George Chiang MD Lynn Byrd MD James Perry MD Amber Hildreth DO Jenni Friedman MD Jonathan Sebat PhD Tina Chambers PhD Albert Oriol Jeff Neul MD PhD Cyndi Kuelbs MD Lauge Farnaes MD PhD Bruce Barshop MD

Kevin Hall PhD James Richardson Kyle Farh MD Susan T

• usi

Severine Catreux Mike Ruehle John Reynders T

• dd Laird

Yan Ding MD Joe Gleeson MD PhD Michelle Clark PhD Julie Cakici RN Wendy Benson Ray Veeraraghavan PhD Matt Bainbrindge PhD Jennifer Azares Sergei Batalov PhD Vanessa Wertheim RN PhD 4 new folk Shareef Nahas PhD FACMG Shimul Chowdhury PhD FACMG David Dimmock MD FACMG Julie Reinke Zornita Stark MD Sue White MD Tiong T an MD

The children are waiting….