Next-Generation Sequencing Exomes are a powerful assay used in both clinical and research settings to discover novel and rare small variants. Now a mature part of many labs, exomes consistently provide coverage over hundreds of thousands of targets across the genome. Along with the small variants, exomes can also be used to call Copy Number Variations, providing extra value for data you may already have and discovering events that may not be captured by any of your existing testing technology. In this webinar, we will address common questions about calling large events on exome data, including: To what extent can Exomes replace Chromosomal MicroArrays (CMAs) for calling large Copy Number Variations (CNVs) What is the validation strategy for a CNV calling method that finds everything from single-exon events to whole chromosome aneuploidy? How Loss of Homozygosity and Copy Number calls be integrated into one analysis and interpretation workflow Watch as we review the next generation CNV and LOH calling algorithm coming to VarSeq and provide case-studies and examples of the capabilities of this algorithm and how it fits into the existing powerful VarSeq platform

1. Calling Large LOH and CNV
Events with NGS Exomes
March 8, 2017
Gabe Rudy
VP Product & Engineering
Dr. Nathan Fortier
Senior Software Engineer
& Field Application Scientist

2. Agenda
Clinical Testing for CNVs: From Small to Large
Roadmap and Special Offer
2
3
4
Method and Demo
Overview Golden Helix1

3. Use the Questions pane in
your GoToWebinar window
Questions during
the presentation

4. Golden Helix – Who We Are
Golden Helix is a global bioinformatics
company founded in 1998.
Filtering and Annotation
Single Sample CNV-Analysis
Clinical Reports
Pipeline: Run Workflows
GWAS
Genomic Prediction
Large-N-Population Studies
RNA-Seq
Large-N CNV-Analysis
Variant Warehouse
Centralized Annotations
Hosted Reports
Sharing and Integration

5. Over 300 customers globally

6. Cited in over 1000 peer-reviewed publications

7. Golden Helix – Who We Are
When you choose a Golden Helix solution, you get more than just software
▪ REPUTATION
▪ TRUST
▪ EXPERIENCE
▪ INDUSTRY FOCUS
▪ THOUGHT
LEADERSHIP
▪ COMMUNITY
▪ TRAINING
▪ SUPPORT
▪ RESPONSIVENESS
▪ TRANSPARENCY
▪ INNOVATION and
SPEED
▪ CUSTOMIZATIONS

8. CNVs in Clinical Testing
▪ Critical evidence needed for many genetic tests
▪ Common driver specific cancers, causal hereditary variation
- EGFR Exon 19 deletion common in lung cancer.
- PIK3CA Amplification in breast cancer
- STK11 deletion associated with Peutz-Jeghers syndrome
- RAI1 deletion associated with Smith-Magenis syndrome
- PTEN gross deletion/duplication ASD, PTEN hamartoma tumor syndrome (PHTS)
▪ Large events used heavily in diagnostics
- Autism Spectrum Disorder (ASD)
- Developmental Delay (DD)
- Intellectual Delay (ID)
- Multiple Congenital Anomalies (MCA).
- Trisomy 21, other trisomy

9. Existing Testing Scenarios for CNV
▪ Clinical Gene Panels
- Test: MLPA, qPCR
- Additionally: high-density microarrays
▪ Clinical Exomes for Germline & Tumor
- Whole-genome micro-arrays (CMAs)
- Lower bound of 10KB event size
- Sensitivity for testing at 100KB
- Known pathogenic CNVs tested with MLPA, qPCR
Nature 459, 569-573 (28 May 2009)
Autism genome-wide copy number variation reveals ubiquitin and neuronal genes

10. Using NGS to Detect CNVs
▪ NGS Being Used Already to Call Small
Variants
- Coverage profile is proportional to input DNA
- With proper normalization, can call CNVs
- As small as single target (exons)
- As large as chromosomal aneuploidy
▪ Validation Requires Multiple Assays
- Samples with known events
- Compare sensitivity / precision vs existing
methods
- Large Events:
- CMAs
- Small Events:
- High-density custom micro-arrays, MLPA,
PCR

11. CNV calling in VarSeq
▪ Reference samples used for
normalization
▪ Metrics
- Z-score: number of standard deviations
from reference sample mean
- Ratio: sample coverage divided by
reference sample mean
- VAF: Variant Allele Frequency
▪ Probabilistic model used to call CNVs

12. VAF
▪ VAF provides supporting evidence
- Values other than 0 or 1 are evidence against het. Deletions
- Values of 2/3 and 1/3 are evidence for duplications

13. Segmentation
▪ Metrics are noisy over large
regions
▪ Outliers cause large events to be
called as many small events
▪ Addressed using segmentation:
- CNAM Optimal Segmentation
- Regions containing many events are
segmented
- Small events sharing a segmented
region are merged

14. QC Flags
▪ Low quality events can be flagged if
- Event targets have low coverage
- There is high variation between samples at event targets
- Event cannot be differentiated from noise at a region
▪ Samples can be flagged if
- The sample does not match the references
- The sample has extremely low coverage
- There is high variance across the target regions

15. Reference Samples
▪ Matched references are chosen for each
sample
▪ Samples with lowest percent difference are
chosen
▪ Performance affected if controls don’t have
matching coverage profile
▪ Samples are flagged if the average percent
difference is above than 20%

16. Experiment
▪ We tested our algorithm on 25
Exomes containing 72 CNV events
called via CMA
▪ An event is considered a false
positive if
- it is called by our algorithm,
- not flagged, and
- larger than 51 kb
▪ An event is considered a false
negative if it is not called by our
algorithm
Results
▪ Most samples had 2 or fewer false
positives
▪ Only one sample had more than 4
false positives
Tests on Exome Data

17. Preliminary Algorithm Comparison
▪ We compared our approach to
CoNVaDING which uses thresholding
on the Z-Score and Ratio
▪ CoNVaDING called very few false
positives but failed to call around half
of the events
▪ Multi-gene and whole chromosome
events were called as collections of
small events
Sensitivity Precision
Golden
Helix
89 % 66 %
CoNVaDING 54 % 99 %

18. Requirements
▪ 100x Coverage
▪ Reference samples
- Recommend at least 30 references
- Minimum of 10
- From same platform and library
preparation
- Gender matched references required
for Non-autosomal calls

19. LoH Calling
▪ Large LoH events need to be
interpreted in any gene test that
covers large CNVs
▪ New Loss of Heterozygosity(LOH)
detection based on H3M2 (Magi et al.)
▪ Calls LoH events using Hidden
Markov Model (HMM)
- Observations are variant allele frequencies
- States are either Homozygous or Non-
Homozygous
▪ Outperforms competing algorithm
(GERMLINE and PLINK) when
compared on the same dataset

20. LoH Calling

21. VarSeq Demonstration

22. Roadmap
▪ Exome Analysis
- Mid-March VarSeq 1.4.4
- Add-on Feature
▪ CNV Reporting
- Able to add to CNVs to VSReport
▪ CNV Annotations
- CNV annotations currently available
- Next release integrate regional/overlap
annotation of CNVs
- Allow for more advanced filtering and
interpretation workflows

23. VarSeq with CNV
▪ 1 seat of VarSeq with CNV analysis - $9,995!
or
▪ 1 seat of VarSeq with CNV analysis + VS Reports, OMIM &
CADD - $14,995!
▪ 15 month license
▪ Offer expires on April 30, 2017 and is for for new customers
only
▪ Request a personal demo by emailing us
info@goldenhelix.com
Special Offer

24. Questions or
more info:
▪ Email
info@goldenhelix.com
▪ Request an evaluation of
the software at
www.goldenhelix.com