Targeted Sequencing Panels on
GIAB Benchmark Genomes
Megan Cleveland, Ph.D.
Applied Genetics Group - NIST
Genome in a Bottle Consortium Workshop
September 15, 2016
Purpose - Goals
• To perform targeted sequencing of GiaB reference materials (and
candidates)
• Using commercial assays
• Further characterization of GiaB with assays of interest (clinical
relevance)
• Compare to high confidence regions
• Identify/mitigate any discordant base calls
• Potentially add new high confidence sequence information to the materials
Working in coordination with Justin and Marc
Samples for Targeted Sequencing
NA12878
Ashkenazi Trio
RM 8392
Asian Trio
GM24143GM24149
GM24385
NIST RM 8391
GM24695GM24694
GM24631
NIST RM 8393
Caucasian Female
NIST RM 8398
7 Individuals
3 replicates
Selection Targeted Sequencing Panels
Ion AmpliSeq™ Pharmacogenomics Panel
Ion AmpliSeq™ Cancer Hotspot Panel v2
Ion AmpliSeq™ Comprehensive Cancer Panel
Ion AmpliSeq™ Inherited Disease Panel
Ion AmpliSeq™ Exome Panel
HID-Ion AmpliSeq™ Identity Panel
HID-Ion AmpliSeq™ Ancestry Panel
TruSight One
TruSight HLA
TruSight RNA Pan-Cancer
TruSight Myeloid
TruSight Cancer
TruSight Tumor 15
TruSight Cardio
TruSight Inherited Disease
A starting point – input is welcome!
Overlap between TruSight and AmpliSeq IDPs
TruSight Inherited Disease Panel
550 genes
AmpliSeq Inherited Disease Panel
328 genes
454 96 232
Regions of interest are captured by
hybridization to biotinylated probes
and isolated by magnetic pulldown.
Regions of interest are amplified using
>10,000 primer pairs
Comparing the Inherited Disease Panels to
the GIAB truth set
Data Analysis
Sequence
Replicate 1
IDP VCF
Generate VCF with
corresponding
commercial Software
GIAB
VCF
IDP
VCF
Compare the VCF files
(ideal would be 100%
overlap)
Sequence
Replicate 2 IDP VCF
Generate VCF with
corresponding
commercial Software
GIAB
VCF
IDP
VCF
Compare the VCF files
(ideal would be 100%
overlap)
Sequence
Replicate 3 IDP VCF
Generate VCF with
corresponding
commercial Software
GIAB
VCF
IDP
VCF
Compare the VCF files
(ideal would be 100%
overlap)
VCF from AmpliSeq or TruSight IDP
GIAB VCF
Intersection of GIAB and IDP bed files
GIAB
Truth
bed file
Inherited
Disease
Panel
bed file
New
Bed
File
Intersection of the Two Bed Files
~90% of each IDP
covered in GIAB
bed file
Examine Discordant Results in IGV
• Vcfeval + Hap.py reports false negatives and false positives, with
respect to the GIAB VCF
• The alignments in those locations are examined in IGV to determine
why the result is discordant
Source: broadinstitute.org
We welcome input regarding:
• Additional panels
• Data analysis
• Other
Megan.Cleveland@nist.gov
Acknowledgements
• Applied Genetics Group
• Pete Vallone
• Kevin Kiesler
• Lisa Borsuk
• Genome Scale Measurements Group
• Justin Zook
• Marc Salit
• Jennifer McDaniel
• David Catoe
• Lindsay Vang
Please email me with any
questions or comments
Megan.Cleveland@nist.gov

GIAB Sep2016 Lightning megan cleveland targeted seq

  • 1.
    Targeted Sequencing Panelson GIAB Benchmark Genomes Megan Cleveland, Ph.D. Applied Genetics Group - NIST Genome in a Bottle Consortium Workshop September 15, 2016
  • 2.
    Purpose - Goals •To perform targeted sequencing of GiaB reference materials (and candidates) • Using commercial assays • Further characterization of GiaB with assays of interest (clinical relevance) • Compare to high confidence regions • Identify/mitigate any discordant base calls • Potentially add new high confidence sequence information to the materials Working in coordination with Justin and Marc
  • 3.
    Samples for TargetedSequencing NA12878 Ashkenazi Trio RM 8392 Asian Trio GM24143GM24149 GM24385 NIST RM 8391 GM24695GM24694 GM24631 NIST RM 8393 Caucasian Female NIST RM 8398 7 Individuals 3 replicates
  • 4.
    Selection Targeted SequencingPanels Ion AmpliSeq™ Pharmacogenomics Panel Ion AmpliSeq™ Cancer Hotspot Panel v2 Ion AmpliSeq™ Comprehensive Cancer Panel Ion AmpliSeq™ Inherited Disease Panel Ion AmpliSeq™ Exome Panel HID-Ion AmpliSeq™ Identity Panel HID-Ion AmpliSeq™ Ancestry Panel TruSight One TruSight HLA TruSight RNA Pan-Cancer TruSight Myeloid TruSight Cancer TruSight Tumor 15 TruSight Cardio TruSight Inherited Disease A starting point – input is welcome!
  • 5.
    Overlap between TruSightand AmpliSeq IDPs TruSight Inherited Disease Panel 550 genes AmpliSeq Inherited Disease Panel 328 genes 454 96 232 Regions of interest are captured by hybridization to biotinylated probes and isolated by magnetic pulldown. Regions of interest are amplified using >10,000 primer pairs
  • 6.
    Comparing the InheritedDisease Panels to the GIAB truth set
  • 7.
    Data Analysis Sequence Replicate 1 IDPVCF Generate VCF with corresponding commercial Software GIAB VCF IDP VCF Compare the VCF files (ideal would be 100% overlap) Sequence Replicate 2 IDP VCF Generate VCF with corresponding commercial Software GIAB VCF IDP VCF Compare the VCF files (ideal would be 100% overlap) Sequence Replicate 3 IDP VCF Generate VCF with corresponding commercial Software GIAB VCF IDP VCF Compare the VCF files (ideal would be 100% overlap)
  • 8.
    VCF from AmpliSeqor TruSight IDP GIAB VCF Intersection of GIAB and IDP bed files
  • 9.
    GIAB Truth bed file Inherited Disease Panel bed file New Bed File Intersectionof the Two Bed Files ~90% of each IDP covered in GIAB bed file
  • 10.
    Examine Discordant Resultsin IGV • Vcfeval + Hap.py reports false negatives and false positives, with respect to the GIAB VCF • The alignments in those locations are examined in IGV to determine why the result is discordant Source: broadinstitute.org
  • 11.
    We welcome inputregarding: • Additional panels • Data analysis • Other Megan.Cleveland@nist.gov
  • 12.
    Acknowledgements • Applied GeneticsGroup • Pete Vallone • Kevin Kiesler • Lisa Borsuk • Genome Scale Measurements Group • Justin Zook • Marc Salit • Jennifer McDaniel • David Catoe • Lindsay Vang Please email me with any questions or comments Megan.Cleveland@nist.gov