An update on the progress of the BRCA Challenge and the BRCA Exchange

1. PROGRESS AND PLANS
Gunnar Rätsch, Ph.D.
Memorial Sloan Kettering Cancer Center, New York, USA
ETH Zürich, Switzerland (since May 2016)
THE BRCA CHALLENGE & EXCHANGE

2. Goals of the BRCA Challenge
To improve the care of patients at risk of breast and ovarian cancer
using global data sharing and collaboration in the analysis of BRCA1
and BRCA2
1. Share BRCA1 and BRCA2 variants publically via a web portal
1. Displays a curated list of BRCA variants, interpreted by expert
consensus, to enable, without dictating, accurate clinical care
2. Includes an environment for collaborative variant curation with access to
evidence (e.g. phenotypes, family history, genetic data, and functional
studies)
2. Address the social, ethical, and legal challenges to global data
sharing
3. Create a model for all disease genes

3. BRCA
Challenge
Steering
Committee
Data Collection
and
Interpretation
Subcommittee
Evidence
Gathering
Group
Classified
Variant
Collection
Group
Interpretation
Group
Ethico-Legal
and
Engagement
Subcommittee
BRCA Challenge Organization
Gunnar Rätsch, MSKCC/ETH
Antonis Antoniou, CRUK
Heidi Rehm, Harvard
Johan den Dunnen, LUMC
Amanda Spurdle, QIMR
Fergus Couch, Mayo
Bob Cook-Deegan, Duke
Kazuto Kato, UKyoto
Delyth Morgan, Br. Cancer Now

4. Goals of the Challenge
To improve the care of patients at risk of breast and ovarian cancer
using global data sharing and collaboration in the analysis of BRCA1
and BRCA2
1. Share BRCA1 and BRCA2 variants publically via a web portal:
The BRCA Exchange
1. Displays a curated list of BRCA variants, interpreted by expert
consensus, to enable, without dictating, accurate clinical care
2. Includes an environment for collaborative variant curation with access to
evidence (e.g. phenotypes, family history, genetic data, and functional
studies)
2. Address the social, ethical, and legal challenges to global data
sharing
3. Create a model for all disease genes

5. BRCA Exchange Public Site (Tier 1)
• Intended for clinicians, genetics counselors, lay public.
• Show classifications only for expert panel reviewed variants
• List other variants as “not yet classified”, more classifications come soon.
• Include variants from Clinvar, 1000G, BIC, LOVD, ENIGMA, ESP,
exLOVD.
• Links back to databases that contain the variant (e.g., ClinVar, exLOVD).
• All data is public.
Gene/Classifica
tion
Pathogenic Benign Not yet
classified
BRCA1 13 483 7,230
BRCA2 31 502 5,241

6. Each repository contributes distinct
BRCA variations
ClinVar 1000G ExAC BIC LOVD ENIGMA ESP exLOVD
ClinVar 6959 1102 1708 3354 1950 1030 533 285
1000G 4348 562 382 295 820 247 87
ExAC 3737 1019 1049 220 600 205
BIC 3606 1581 274 382 266
LOVD 2993 301 420 283
ENIGMA 1030 146 208
ESP 687 116
exLOVD 294
Combined, BRCA Exchange has 13,500 individual,
deduplicated variants.

7. Different RepositoriesShare Same Variants Reported
Differently
Two simple methods of variant comparison:
Edit variant description Genome Sequence
Repository 1 Variant after reference position 1 “GC” is
inserted in this variant
AGCGCGC
Repository 2 Variant after reference position 5 “GC” is
inserted in this variant
AGCGCGC
String Comparison
says same variant
Direct Comparison
says different variant
Keep track of different variant descriptions on the site.

8. Each repository contributes distinct
information on BRCA variation
Combined, BRCA Exchange soon has
15,000 individual deduplicated variants.

9. brcaexchange.org

10. (1)
(2)
(3)
• Simple user experience:
• (1) Enter variant
• (2) Browse related
variants
• (3) Find out more and
explore linked resources

11. Clinical significance calls only following expert review by ENIGMA

12. Goals of the BRCA Challenge
To improve the care of patients at risk of breast and ovarian cancer
using global data sharing and collaboration in the analysis of BRCA1
and BRCA2
1. Share BRCA1 and BRCA2 variants publically via a web portal:
The BRCA Exchange
1. Displays a curated list of BRCA variants, interpreted by expert
consensus, to enable, without dictating, accurate clinical care
2. Includes an environment for collaborative variant curation with access to
evidence (e.g. phenotypes, family history, genetic data, and functional
studies)
2. Address the social, ethical, and legal challenges to global data
sharing
3. Create a model for all disease genes

13. BRCA Exchange for Research (Tier 2 & 3)
Main Purpose
Support and enable variant classification by aggregating relevant public and
controlled access data in two tiers and serve as data exchange hub to receive and
distribute data submissions
Tier 2 (open access, released in April 2016)
• Public access, but site is available through a disclaimer
• Show all available classifications from different sources
• Provide additional information sources such as population frequencies,
computational predictions, etc.
• Provide tools to identify disagreements (soon: support & track curation)
Tier 3 (controlled access, ongoing developments)
• Links to case level & detailed clinical information and other restricted data
• Controlled access: Develop approval mechanism for continuous data flow from
data sources to approved research groups
• Develops input interfaces with case level data providers lab test companies,
research institutions etc. who share data for research
• Provide computational tools to support variant classifications by experts

14. One can download all data
Clinical significance calls from original submitters (including conflicting calls)

15. Filters allow increased personalization of the research data

16. Lollipops show variant distribution along the length of each gene

17. BRCA Exchange is already used by
curators and the public
BRCA Exchange already attracts ≈1,000 unique users per month.
(currently ≈100 page views per day).

18. Future Plans
• More data sources! Plan to include TCGA, Quest & Haplotype Research
Consortium data.
• Contact me (raetsch@ethz.ch) or Rachel Liao
(rachel.liao@genomicsandhealth.org) if you have variants or case level
data that we can include.
• Currently data is imported to central site.
Plan: Implement full GA4GH
data federation for variant
data sources
• Improve User Interface
• Semi-automatic variant
assessment

19. Currently, curators manually amass all
data on a variant ...
Variant
Classification
Cosegregation
Pathology
Probability
calculationMore ...
Physiochemic
al properties
(A-GVGD
Score)
Co-
occurrence
Family
history
Slow
Cumbersome
Requires great attention to
prevent errors

20. BRCA
Exchange
Research
Variant
Classification
Cosegrega
tion
Pathology
Probability
calculation
More ...
Physiochemica
l properties (A-
GVGD Score)
Co-
occurrence
Family
history
BRCA Exchange helps curators
➢ Near Future: Expand Research pages by (a) hosting
more data and (b) pre-computing as much as possible

21. Acknowledgements
Molly Zhang
Charlie Markello
Benedict Paten
Mary Goldman
Melissa Cline
Brian Craft
Gunnar Rätsch
Rachel Liao
BRCA Challenge
* Steering Committee
* Evidence Gathering Group
* Interpretation Group
BRCA Challenge Chairs: John Burn and Stephen Chanock

22. Thanks!
• BRCA Challenge:
• John Burn
• Stephen Chanock
• Antonis Antoniou
• Larry Brody
• Robert Cook-Deegan
• Fergus Couch
• Johan den Dunnen
• Susan Domchek
• Douglas Easton
• William Foulkes
• Judy Garber
• David Goldgar,
• Kazuto Kato
• Delyth Jane Morgan
• Robert Nussbaum
• Kenneth Offit
• Sharon Plon
• Gunnar Rätsch
• Nazneen Rahman
• Heidi Rehm
• Mark Robson
• Wendy Rubinstein
• Amanda Spurdle
• Dominique Stoppa-Lyonnet
• Sean Tavtigian
• David Haussler
BRCA-Exchange.org:
• Molly Zhang
• Brian Craft
• Charles Markello
• Mary Goldman
• Jing Zhu
• Melissa Cline
• Benedict Paten
• Gunnar Rätsch
• Rachel Liao
• GA4GH Clinical Work Group
• Peter Goodhand
• (And many more!)

23. ClinVar
LOVD
Case data and
evidence sharing
BRCA
Exchange
Research
National nodes
Commercial &
Research
Population
Frequency Data
(EVS, ICGC,
1000G, 10KUK)
Functional
Data
Segregation Co-occurrence
In silico
prediction
Options for variant
submission
BRCA
Exchange
http://brcaexchange.orgIG
Variant
Classification
BRCA Research Data & Curation
Community

24. BRCA Challenge Steering Committee
25
Sir John Burn, Newcastle University (United Kingdom) – Co-Chair
Stephen Chanock, National Cancer Institute (United States) – Co-Chair
Antonis Antoniou, University of Cambridge (United Kingdom)
Larry Brody, National Human Genome Research Institute (United States)
Robert Cook-Deegan, Duke University (United States)
Fergus Couch, Mayo Clinic (United States)
Johan den Dunnen, Leiden University Medical Center (Netherlands)
Susan Domchek, University of Pennsylvania (United States)
Douglas Easton, University of Cambridge (United Kingdom)
William Foulkes, McGill University (Canada)
Judy Garber, Dana Farber Cancer Institute (United States)
David Golgar, Huntsman Cancer Center (United States)
Kazuto Kato, Osaka University (Japan)
Baroness Delyth Morgan, Breast Cancer Now (United Kingdom)
Robert Nussbaum, Invitae (United States)
Ken Offit, Memorial Sloan Kettering Cancer Center (United States)
Sharon Plon, Baylor College of Medicine (United States)
Nazneen Rahman, Institute of Cancer Research (United Kingdom)
Gunnar Rätsch, Memorial Sloan Kettering Cancer Center (United States)
Heidi Rehm, Harvard Medical School (United States)
Mark Robson, Memorial Sloan Kettering Cancer Center (United States)
Wendy Rubinstein, National Institute of Health (United States)
Amanda Spurdle, QIMR Berghofer Medical Research Institute (Australia)
Dominique Stoppa-Lyonnet, Curie Institute (France)
Sean Tavtigian, University of Utah (United States) Underline denotes leadership
on a subgroup

25. Motivation for the BRCA Exchange
● ClinVar is incomplete:
○ European projects
○ Individual papers and submitters
○ Some organizations can't pay the
cost of preparing a submission
BRCA variation is relatively common
with well known medical implications
No single source for BRCA variant
information
ClinVar
7961 Variants
1041
2107
1191
1778
LOVD
3276 Variants
UMD
3675 Variants

26. Motivation for the BRCA Exchange
BRCA variation is relatively common
with well known medical
implications
There is no single source for BRCA
variant information
High numbers of VUSs, where
classification is limited by being
unable to see all the data on a
variant

27. Each repository contributes distinct
information on BRCA variation
Different databases
contribute different
types of information
Database Origin Allele
Frequenc
y
Significan
ce
Publicatio
ns
ClinVar Y N Y N
1000
Genomes
N Y N N
ExAC N Y N N
BIC Y N Y Y
LOVD Y Y Y N
ENIGMA Y N Y Y
ESP N Y N N
exLOVD N N N Y

28. (BEGIN VIDEO)

29. Community Engagement
Involved patient advocacy groups
• FORCE, BRCActivist, Brave Bosom
Community space on website
• Allow networking, especially for people in developing countries
• Newsletter with updates on the BRCA Challenge and how to share
data
Case-level data
• Allow people to donate their own data, similar to 1000 Genomes
• Due to legal challenges, start with USA, then include more
countries
Coming soon ...