Personalized genomics may be moving into a new era with whole-exome and whole-genome sequencing becoming affordable and available to consumers. 23andMe recently piloted a more affordable 80x exome to their existing customers. But it remains to be seen whether this wealth of raw genomic data can be analyzed to provide meaningful results for both healthy and symptomatic individuals.
By acquiring 23andMe exomes on his family, Gabe puts himself in the position of a bioinformatically inclined consumer, but non-clinician, to approach this question with his own analysis. His trio consists of a healthy father and son, and a mother with clinically-diagnosed idiopathic rheumatoid arthritis.
The following goals were set for the analysis: 1) How accurate are variant calls from direct-to-consumer NGS services? 2) How useful and durable is the list of risk variants provided by 23andMe? 3) Can a healthy individual's exome provide additional risk information over standard genotype-array-based risk prediction? and 4) Can the state of our current understanding of the complex genomic architecture of autoimmune diseases be enough to to find potential driver variants and genes to explain the diagnosis of a single case?
Here Gabe presents his findings of this analysis and discuss how individualized genomics might change in the world of affordable sequencing.
Semelhante a AGBT 2013: Home Brewed Personalized Genomics - The Quest for Meaningful Analysis Results of a 23andMe Exome Pilot Trio of Myself, Wife, and Son
Aug2013 Heidi Rehm integrating large scale sequencing into clinical practiceGenomeInABottle
Semelhante a AGBT 2013: Home Brewed Personalized Genomics - The Quest for Meaningful Analysis Results of a 23andMe Exome Pilot Trio of Myself, Wife, and Son (20)
AGBT 2013: Home Brewed Personalized Genomics - The Quest for Meaningful Analysis Results of a 23andMe Exome Pilot Trio of Myself, Wife, and Son
1. The Quest for Meaningful
Analysis Results of a
23andMe Exome Pilot Trio of
Myself, Wife, and Son
February 22, 2013
Gabe Rudy, Vice President of
Product Development
Home Brewed Personalized Genomics
2. Exome Sequencing in Consumer Genomics
Exomes done as part of Pilot
Program
80x coverage
Raw data with no interpretation
Erin
JIA
Gabe
(me)
Ethan
3. Overview
Consumer genetics data: research or clinical grade?
Using exome data to explain a rare autoimmune disorder
1
2
3
Treating my healthy self to a Mendelian disease analysis
4. Consumer exomes: done using best practices
Sequencing done on HiSeq 2000
- 75bp PE
- Agilent SureSelect exome capture
Aligned and called with BWA/GATK
- Broad’s Best Practices with GATK Guide
- Indel realignment
- UnifiedGenotyper called samples concurrently
Deliverables
- BAM (minus indel realignments)
- VCF (some filters applied)
- PDF of Summary Report
5. Research or clinical grade?
Total Reads 140M
Unique Align 87%
Mean Target 105x
% Target at 2x 97%
% Target at 10x 94%
% Target at 20x 89%
% Target at 30x 83%
8. Updated VCF and report at end of October
GATK is a Research Tool. Clinics Beware.
9. Using exome data to explain a rare autoimmune disorder
Treating my healthy self to a Mendelian disease analysis
Overview
1
2
3
Consumer genetics data: research or clinical grade?
Rare Variant Analysis:
The Hammer
My Exome:
The Nail
10. Filtering and analysis strategy
Everything
Low quality &
phantom
Non-synonymous
or
LoF
Rare
1,698
17K
85K
151K
Now what?
Follow best
practices for high-
impact variants
Weed out false-
positives
Use functional
prediction
Interpretation more
open-ended
12. More filtering strategies
Regions of Chromosomal Duplication (SuperDups)
Look at genes in OMIM (most)
Use predictions of genes as recessive/haploinsufficient to
weed out low-priority genes
For nonsynonymous missense variants can use functional
prediction (SIFT/Polyphen2) to annotate
16. Overview
Consumer genetics data: research or clinical grade?1
2
3
Treating my healthy self to a Mendelian disease analysis
Using exome data to explain a rare autoimmune disorder
17. Juvenile Idiopathic Arthritis (JIA)
Unknown cause, onset before 16
Between 8 and 150 of every 100,000
children
50% have pauciarticular JIA
40% have polyarticular JIA
Polyarticular RF negative sub-phenotype
has heritability similar to Rheumatoid
Arthritis (RA)
- RA is expected to be 60% heritable
- 51% explained through current genetic
associations
- 36% of heritability in HLA
18. IVA – Rare deleterious variance within 1 hop of JIA
30. No smoking gun, but variants of interest
Ongoing RA research with population level WGS and family NGS
To be seen how much of autoimmune disorder heritability is explained by rare
variants with higher effect sizes.
Most promising signal is in genotype SNPs that might be tagging for functional
mutations in regulatory regions.
Rare sub-classifications like JIA polyarticular RF negative
may be difficult to nail down with population studies
Family studies looking at shared biomarkers along
with symptoms may be better suited to find cause-effect
relationships
Wife’s nuclear family has diagnosed cases of:
̵ Pheochromocytoma: 2-8 per 1,000,000
̵ Guillain-Barré: 0.6-4 per 100,000
Final thoughts
31. Acknowledgements
Dr. Peter Gregersen
- Director, Robert S. Boas Center for Genomics and Human
Genetics, The Feinstein Institute
Dr. Gerald Nepom
- Director, Benaroya Research Institute, Director, Immune
Tolerance Network
Golden Helix
- SNP & Variation Suite
- GenomeBrowse
Sean Scott - Ingenuity
- Variant Analysis
Tim Hague - Omixon
- Target HLA Typing
Dr. Brian Naughton, 23andMe
- Trio exome sequencing
Dr. Peter Gregersen
Dr. Gerald Nepom
Notas do Editor
Show these in GB more than talking about them here
OMIM or Recessive technically
Known or Likely Recessive P(Rec) > 0.7 or in a gene with a known recessive pathogenic pattern
Not in NHLBI, 1KG, has pathogenic allele status in dbSNP for OTC
Deficiency of proper OTC production can lead to metabolic crisis due to ammonia accumulation (especially in terms of protein breakdown)
Such low penetrance of gene variants is particularly common in conditions affecting nutrient metabolism.
23andMe Provides Genotyping Service
~1M SNPs genotyped
48 Diseases Carrier Status
57 Traits
20 Drug Responses
119 Diseases Risk Predictions
Exome done as Pilot Program
80X coverage
Raw Data
No Interpretation
Carrier Status: Cystic Fibrosis, some BRCA mutations
Traits: Blood Glucose/LDL Cholesterol Levels to Hair Thikness and Freckling to
Drug Responses: - Warfarin, Statin response. Caffein Metabolism and various effacacy and toxisity PGX
Disease Risk: Various Specific Cancers, Celiacs, Parkinson’s, Alzheimer’s (Age Related Macular Degeneration – my 3x increased risk)
Narcisism
Some interesting rare variants here
Need more notes no thoughts here
Encrypted bundle
Have 677 poly RF- in CCHMC study through
Maybe put # JIA and # Poly RF-