Next Generation Sequencing as a Tool for Cancer

Next Generation
Sequencing as a Tool
for Cancer Precision
Medicine
Karina Eterovic, PhD
Assistant Professor
Institute for Personalized
Cancer Therapy (IPCT)
Rio de Janeiro, April 2015
T200 – A Targeted Sequencing Platform
developed by the IPCT laboratory
DNA
Bases
T200
All exons in 201 cancer related genes
1 Mb
At minimum 500 x (average 800-950x)
Whole Exome
Every single gene – EXONS ONLY
Whole Genome
Every single gene – FULL LENGHT
60 Mb
3 Gb
T200 vs Whole Exome
Tumors are
heterogeneous
T200 vs Whole Exome
Tumors are
heterogeneous
T200T200
genes
IPCT
Panel
•
Mutation prevalence (COSMIC and TCGA)
- 3% or more samples in all diseases combined
- 4-5% samples in a specific disease
- Minimum of 50 samples analyzed
•
•
Actionable: drug in trial or in the pipeline
MDA faculty input
5
Our Study - Samples
Clearing House Protocol PA11-0852
PI – Funda Meric-Bernstam
Patients samples (475 patients)
FFPE tumor and matched normal (blood)
Goals:
1. To determine feasibility of implementing standardize genomic testing
2. To determine the frequency of mutations and co-mutations
3. To establish a database
4. To determine how we can predict response
Currently: over 5000 samples processed
T200 Workflow
T200 Workflow
T200 Workflow
Ken Chen, PhD
How much sequencing depth is necessary
for clinical samples?
10%
Sanger
sensitivity
40x
*If DNA fragmentation is random, duplicated reads are removed, and both the reference and the variant alleles are
equally represented: the number of reads that are sampled from a variant allele follows a binomial distribution
How much sequencing depth is necessary
for clinical samples?
5%
100x
*If DNA fragmentation is random, duplicated reads are removed, and both the reference and the variant alleles are
equally represented: the number of reads that are sampled from a variant allele follows a binomial distribution
How much sequencing depth is necessary
for clinical samples?
Adaptive threshold
1%
700x
Not cost-effectively feasible by WGS or WES as
currently implemented
Concordance between fresh frozen and FFPE
Cell lines (fresh frozen tissue)
48.5 kb
Concordance between fresh frozen and FFPE
Cell lines (fresh frozen tissue)
FFPE tissue
- Double strand breaks
48.5 kb
- Hydrolysis of N-Glycosyl Bonds
- Base Deamination (dC to dU)
- Oxidation
Low library yield or failure
Sequencing errors
Concordance between fresh frozen and FFPE
Over 500 reads
r2= 0.9445
Concordance between fresh frozen and FFPE
Over 500 reads
r2= 0.9445
200-499 reads
r2= 0.9208
Concordance between fresh frozen and FFPE
Over 500 reads
r2= 0.9445
200-499 reads
r2= 0.9208
Less than 200 reads
r2= 0.8022
Concordance between fresh frozen and FFPE
Over 500 reads
r2= 0.9445
200-499 reads
r2= 0.9208
Less than 200 reads
r2= 0.8022
Does the amount of input DNA matter?
Duplicates
Coverage
Does the amount of input DNA matter?
Duplicates
T200 cutoff: 150 ng – standard protocol
50 ng – low input protocol
Coverage
Disease Sites
Coverage
Disease Sites
Coverage
Disease Sites
Coverage
T200 Genomic Landscape
T200 Genomic Landscape
T200 Genomic Landscape
1-9% MAF
Primary vs Metastasis Heterogeneity
Low MAF mutations detected
exclusively in the primary or
metastatic tumor
Clinical Actionability
Would be missed by WGS or WES as currently implemented
Clinical Actionability
Would be missed by WGS or WES as currently implemented
T200 Validation
T200 Validation
Known mutations
T200 Validation
Known mutations
NEW mutations (also 98% concordance)
T200.1
2.4 Mb
263 genes frequently mutated in cancer
global copy number
Fusion will be added soon (T200.2)
Other supportive IPCT platforms
IPCT Precision Oncology Decision Support (PODS)
List of all genotype-selected and genotype-relevant clinical trials
2927 cancer clinical trials annotated
All of MDA open trials are annotated continuously
Continuously expanding website; Information updated on a routine-basis
PersonalizedCancerTherapy.org
Functional Genomics Platform
Drug screening with cell lines (BAF3_IL-3 dependent) with introduced
genomic aberrations (mutations, fusions)
IPCT sequencing platforms
3. Droplet digital PCR
1. Illumina pipeline
One mutation at the time
Taqman probes (known mutations)
VERY high sensitivity
T200 and T200.1
Whole Exome
RNA sequencing
QX200
HiSeq
2. Ion Torrent pipeline
Ampliseq50 and 400 (CMS50 and 400)
Amplicon sequencing
PGM
4. Oncoscan Affymetrix
Global Copy Number
Few mutations
Oncoscan
Acknowledgements
IPCT Leadership
John Mendelsohn
Gordon Mills
Funda Meric-Bernstam
Kenna M Shaw
My team in the IPCT Lab
Qingxiu Zhang
Nader Ezzeddine
Yu-ye Wen
Maribel Mosqueda
Lin-ya Tang
Ping Song
Ji Wang
MDA Sequencing CCSG Core (SMF)
All MDA departments that provided patient
samples
Data Analysis team
Ken Chen
Hao Zhao
Tae-Beom Kim
Xiaofeng Zheng
Precision Oncology
Decision Support (PODS)
Amber Johnson
Ann Bailey
Vijay Holla
Jia Zeng
Beate Litzenburger