The study of recurrent somatic alterations associated with pediatric, childhood and young adult cancers has lagged behind those that associated with adult cancers. Whole exome and transcriptome approaches are still being used to support discovery efforts, consequently, due to several initiatives aimed at profiling genomic alterations associated with childhood cancers, a set of recurrent somatic alterations has been defined.

1. For Research Use Only. Not for use in diagnostic procedures. Thermo Fisher Scientific • 110 Miller Ave • Ann Arbor, MI 48104 • thermofisher.com
Table 2. Confirmed high-level copy-number amplifications
INTRODUCTION
The study of recurrent somatic alterations associated with pediatric, childhood and
young adult cancers has lagged behind those that associated with adult cancers.
Whole exome and transcriptome approaches are still being used to support
discovery efforts, consequently, due to several initiatives aimed at profiling genomic
alterations associated with childhood cancers, a set of recurrent somatic alterations
has been defined. One example - the Children’s Oncology Group - is actively
pursuing an national effort to characterize childhood cancers and match them to
available targeted therapies through the Pediatric MATCH initiative. Since childhood
cancers are diverse and relatively rare, it is essential to profile both solid tumor and
hematological tissue types, supporting samples from FNA, FFPE, bone marrow, and
blood. To aid in the reporting of the variants derived from such a comprehensive
assay, a full informatics processing workflow is key to focus the variant lists to the
likely relevant alterations. To accelerate research in this area, we have developed a
novel targeted next-generation sequencing assay to detect relevant somatic
alterations previously reported in these cancer types, including translocations,
somatic mutations, and copy number amplifications.
METHODS
The assay was developed using Ion AmpliSeq™ targeted sequencing technology to
provide accurate targeted detection of both DNA and RNA variants from as little as
10ng of input nucleic acid per amplicon pool. Two pools of DNA and two pools of
RNA amplicons were designed, and include 3,060 and 1,704 targeted regions,
respectively. Average length of amplicon inserts is ~120 bp for DNA and ~100 bp for
RNA. 202 unique gene targets were included, of which 130 genes are assayed for
mutations, 28 genes for copy number, and 88 fusion driver genes. Variant calling
algorithms for both DNA and RNA were optimized and combined into a single Ion
Reporter workflow.
To assess SNV and Indel calling performance the AcroMetrix™ Oncology Hotspot
Control was utilized. It was diluted (in the background of gDNA) to assess detection
at 5%, 10%, and 15% frequencies. Amplifications were assessed on a set of 9 FFPE
samples orthogonally profiled with FISH. Pediatric bone marrow samples were
assayed for orthogonally characterized hematological translocations. Additional
pediatric FFPE samples and SeraCare Seraseq™ FFPE v2 , UHR, and HD784
controls were used to assess fusion variants representative of solid tumors.
Sample extraction was performed using RecoverAll™. Up to 10ng of nucleic acid per
pool for a total of 20ng DNA and 20ng RNA was used as input material per sample.
Library preparation was performed manually using AmpliSeq™ Plus or automated
using Ion™ Chef, and all templating was performed on the Ion™ Chef. A total of 7
DNA and 7 RNA samples were run on the same chip at a 80:20 library ratio, with
DNA and RNA negative controls.
RESULTS
The Oncomine™ Childhood Cancer Research Assay (OCCRA) is based on
AmpliSeq™ sequencing technology, Ion S5™ sequencing and Oncomine™
informatics workflow. The assay generated an average read depth of >2,500 reads
per DNA amplicon with high uniformity (>95%), when up to 7 sample DNA-RNA pairs
were analyzed with the 540 chip on Ion S5™ sequencing instrument. Minimal allele
frequency detected for key hotspots was 5%. Sensitive detection of CNV and fusion
variants associated with pediatric solid tumors and hematological cancers was
demonstrated in orthogonally profiled FFPE, blood, and bone marrow samples.
Performance was robust across sample types. Similar results were observed with
manual and automated library preparation.
CONCLUSIONS
A novel NGS assay, designed specifically for research of pediatric, childhood and
young adult cancers, and capable of detecting relevant DNA and RNA alterations from
the same sample, was developed and verified. The assay is useful for characterizing
relevant alterations in a wide range of cancers including childhood leukemias and
lymphomas as well as solid tumors including neuroblastoma, rhabdomyosarcoma,
retinoblastoma, osteosarcoma, Ewing sarcoma, Wilms tumor, brain and spinal cord
tumors.
ACKNOWLEDGEMENTS
The authors express enormous gratitude to Children’s Hospital of Los Angeles for
cooperation in conception and development of the assay, and the patients who have
donated the samples to advance the research of childhood cancers and helped make
the world a healthier place.
Nick Khazanov1, Chaitali Parikh1, Scott P. Myrand1, Efren Ballesteros-Villagrana1, Jingwei Ni1, Paul Williams1, Karen Clyde1, Dinesh Cyanam1, Jon Sherlock1, Jonathan Buckley2, Matthew Hiemenz2, Gigi Ostrow2, Alex Judkins2, Xiaowu Gai2, Tracy Busse2, Alan Wayne2, Deepa Bhojwani2, Gordana Raca2, Matthew
Oberley2, David Parham2, Janice Au-Young1, Jaclyn Biegel2, Seth Sadis1, Timothy Triche2. 1 Thermo Fisher Scientific, Ann Arbor, MI 2 Children’s Hospital Los Angeles, Los Angeles, CA
Development of a next-generation (NGS) assay for pediatric, childhood, and young adult
cancer research, with comprehensive DNA and RNA variant detection
FFPE, blood,
bone marrow
≥ 20 ng RNA/DNA
SuperScript™ IV
VILO™
or Automated
AmpliSeq™
Automated
Ion Chef™
Ion S5 ™
540 chip
Torrent Suite
Ion Reporter
Oncomine Filtering
Table 1. Gene-level content of Oncomine™ Childhood Cancer Research Assay
Hotspot coverage extends to regions of known driver mutations. Full-gene coverage is intended to assay
deleterious mutations in genes that may function as a tumor suppressor. Fusion driver genes only are
listed, with each driver gene often assayed for multiple observed partners. Expression control targets are
also included on the assay (but not listed) to aid with expression normalization.
ABL1
ABL2
ALK
ACVR1
AKT1
ASXL1
ASXL2
BRAF
CALR
CBL
CCND1
CCND3
CCR5
CDK4
CIC
CREBBP
CRLF2
CSF1R
CSF3R
CTNNB1
DAXX
DNMT3A
EGFR
EP300
ERBB2
ERBB3
ERBB4
ESR1
EZH2
FASLG
FBXW7
FGFR1
FGFR2
FGFR3
FLT3
GATA2
GNA11
GNAQ
H3F3A
HDAC9
HIST1H3B
HRAS
IDH1
IDH2
IL7R
JAK1
JAK2
JAK3
KDM4C
KDR
KIT
KRAS
MAP2K1
MAP2K2
MET
MPL
MSH6
MTOR
MYC
MYCN
NCOR2
NOTCH1
NPM1
NRAS
NT5C2
PAX5
PDGFRA
PDGFRB
PIK3CA
PIK3R1
PPM1D
PTPN11
RAF1
RET
RHOA
SETBP1
SETD2
SH2B3
SH2D1A
SMO
STAT3
STAT5B
TERT
TPMT
USP7
ZMYM3
APC
ARID1A
ARID1B
ATRX
CDKN2A
CDKN2B
CEBPA
CHD7
CRLF1
DDX3X
DICER1
EBF1
EED
FAS
GATA1
GATA3
GNA13
ID3
IKZF1
KDM6A
KMT2D
MYOD1
NF1
NF2
PHF6
PRPS1
PSMB5
PTCH1
PTEN
RB1
RUNX1
SMARCA4
SMARCB1
SOCS2
SUFU
SUZ12
TCF3
TET2
TP53
TSC1
TSC2
WHSC1
WT1
XIAP
ALK
ABL2
BRAF
CCND1
CDK4
CDK6
EGFR
ERBB2
ERBB3
FGFR1
FGFR2
FGFR3
FGFR4
GLI1
GLI2
IGF1R
JAK1
JAK2
JAK3
KIT
KRAS
MDM2
MDM4
MET
MYC
MYCN
PDGFRA
PIK3CA
ABL1
ABL2
ALK
BCL11B
BCOR
BCR
BRAF
CAMTA1
CCND1
CIC
CREBBP
CRLF2
CSF1R
DUSP22
EGFR
ETV6
EWSR1
FGFR1
FGFR2
FGFR3
FLT3
FOSB
FUS
GLI1
GLIS2
HMGA2
JAK2
KAT6A
KMT2A
KMT2B
KMT2C
KMT2D
LMO2
MAML2
MAN2B1
MECOM
MEF2D
MET
MKL1
MLLT10
MN1
MYB
MYBL1
MYH11
MYH9
NCOA2
NCOR1
NOTCH1
NOTCH2
NOTCH4
NPM1
NR4A3
NTRK1
NTRK2
NTRK3
NUP214
NUP98
NUTM1
NUTM2B
PAX3
PAX5
PAX7
PDGFB
PDGFRA
PDGFRB
PLAG1
RAF1
RANBP17
RARA
RECK
RELA
RET
ROS1
RUNX1
SS18
SSBP2
STAG2
STAT6
TAL1
TCF3
TFE3
TP63
TSLP
TSPAN4
UBTF
USP6
WHSC1
YAP1
ZMYND11
ZNF384
BCL2
BCL6
FGFR1
FGFR4
IGF1R
MET
MYCN
MYC
TOP2A
HOTSPOT CNV FULL GENE FUSION
EXPRESSION
Figure 2. Mapped reads for DNA samples
Total and amplicon-mapped DNA reads show excellent on-target rates for FFPE, bone marrow, and
control samples. Median of ~8M reads per sample provides >2,500 average read depth, with >95%
average uniformity, providing coverage for detection of low-frequency (≥5% AF) somatic variants.
Figure 3. Distribution of amplicon reads for DNA samples
GC%
Reads
Total reads versus GC% content in 3,060 DNA amplicons on the assay. Shown for all DNA samples
(replicates seen as vertical “stripes”). Colored by amplicon pool. Certain GC-rich amplicons show poor
performance, especially in difficult (low-quality) samples.
Sample ID Cellularity
Known CN Status Novel
Gene FISH NGS Gene OCCRA
365 90% CCND1 13 13.26
1181615b 99% EGFR 52.9 73.84 CDK4 11.77
1185752b 85% MYCN 7.1 4.64
1191227b 40% PIK3CA 6.2 6.53
GLI2 8.13
CCND1 21.75
1193124b 100% CDK6 16.8 7.74
PDGFRA 8.38
KIT 6.7
1196647b 80% PIK3CA 10.5 10.41
1197046b 65% CCND1 6.1 7.25
1198367b 70% MYC 12.1 10.63 CCND1 26.8
1194253b 75% PIK3CA 13.2 12.24
FFPE samples with known copy-number (CN) amplifications of 6 copies or more show concordant CN
estimates when profiled by the NGS assay. Value reported is the lower-bound of the CN estimate range
provided by calling algorithm. Copy-neutral genes were also confirmed. Novel CN variants also detected.
Figure 1. Oncomine™ Childhood Cancer Research Assay Sample-to-Answer Workflow
Figure 4. Mapped reads for RNA samples and proportion of reads by type
a b
TotalMappedReads
%TotalMappedReads
Table 3. Variant detection in DNA control samples
Dilution Ladder
Expected Observed False Pos False Neg Sensitivity PPV
#1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2
Hotspot SNV
AOHC @ 15% 88 88 88 88 0 0 0 0 1 1 1 1
AOHC @ 10% 88 88 88 87 0 0 0 1 1 0.989 1 1
AOHC @ 5% 88 88 88 88 0 0 0 0 1 1 1 1
Hotspot Indel
AOHC @ 15% 4 4 4 4 0 0 0 0 1 1 1 1
AOHC @ 10% 4 4 4 4 0 0 0 0 1 1 1 1
AOHC @ 5% 4 4 4 4 0 0 0 0 1 1 1 1
Excellent sensitivity and PPV for detection of hotspot variants in AOHC at 5%, 10%, and 15% allele
frequency. Two replicates per condition shown. Only one replicate contains one false-negative variant.
a) Total RNA depth provides sensitivity for fusion variants while maintaining robust expression levels.
b) Proportion of expression control reads robust across sample type and quality.
Total Mapped Reads: 829,066 1,717,299 2,351,771 5,241,213
%TMR Reads %TMR Reads %TMR Reads %TMR Reads
CD74-ROS1.C6R34.COSF1200 0.24% 2,010 0.28% 4,888 0.34% 8,089 0.62% 32,471
EGFR-SEPT14.E24S10 2.52% 20,858 2.33% 40,060 3.09% 72,600 3.02% 158,429
EML4-ALK.E13A20.COSF408.1 0.57% 4,765 1.03% 17,719 1.17% 27,450 1.55% 81,283
EML4-ALK.E20A20.COSF409.1 0.05% 857 0.03% 663
EML4-ALK.E6aA20.AB374361 0.00% 45 0.00% 95
ETV6-NTRK3.E5N15.COSF571.1 4.33% 35,923 3.65% 62,755 5.07% 119,243 2.81% 147,343
FGFR3-BAIAP2L1.F17B2.COSF1346 1.41% 11,669 1.57% 27,037 1.64% 38,587 1.96% 102,611
FGFR3-TACC3.F17T10 0.01% 76 0.01% 194 0.14% 3,248 0.04% 1,944
FGFR3-TACC3.F17T11.COSF1348 4.78% 39,602 2.04% 35,101 4.38% 102,910 4.35% 227,820
FGFR3-TACC3.F18T11del5 0.01% 64 0.00% 54 0.00% 70
KIF5B-RET.K24R11.COSF1262 3.29% 27,294 3.87% 66,469 3.66% 86,096 3.94% 206,438
LMNA-NTRK1.L2N11 1.88% 15,607 1.24% 21,368 1.30% 30,539 2.02% 105,632
MYH9-BRD1.M1B2.Non-Targeted 0.03% 1,406
NCOA4-RET.N7R12.COSF1491 1.10% 9,115 0.78% 13,366 1.20% 28,197 1.32% 69,377
SLC34A2-ROS1.S4R34.COSF1198 0.49% 4,067 0.54% 9,269 0.57% 13,452 0.66% 34,530
SLC45A3-BRAF.S1B8.COSF871 1.26% 10,419 1.13% 19,386 1.28% 30,014 1.36% 71,305
SLC45A3-BRAF.S1B8.Non-Targeted 0.13% 6,909
TPM3-NTRK1.T7N10.COSF1329 6.60% 54,685 6.11% 104,941 4.98% 117,001 4.56% 238,782
Table 4. Fusions identified in replicates of Seraseq™ FFPE v2 control .
PRB223: BCR-ABL1 %TMR Read Count
BCR-ABL1.B14A2 5.97% 102,813
ETV6-RUNX1.E5R4 1.54% 26,576
BCR-ABL1.B13A2 0.02% 261
CPM226: KIAA1549-BRAF.K14B9 %TMR Read Count
KIAA1549-BRAF.K14B9.COSF483 0.95% 13,182
CPM278: ETV6-RUNX1 %TMR Read Count
ETV6-RUNX1.E5R3 12.12% 134,367
ETV6-RUNX1.E5R4 2.47% 26,744
EML4-ALK.E20A20.COSF409.1 0.03% 531
P2RY8-CRLF2.P1C1.1 0.02% 151
EML4-ALK.E6aA20.AB374361 0.01% 119
ETV6-RUNX1.E5ins33R3 0.00% 51
P2RY8-CRLF2.P1C2 0.00% 49
The 12 primary isoforms in the SeraCare SeraSeq v2 control were identified reproducibly. Reads also
reported as % of total mapped reads (%TMR) to show relative representation across replicates. Secondary
isoforms and novel non-targeted fusions are observed when total mapped reads are obtained.
Table 5. Fusions identified in clinical samples
PRB212: PML-RARA %TMR Read Count
PML-RARA.P6R3 10.89% 87,643
EML4-ALK.E20A20.COSF409.1 0.84% 8,361
PML-RARA.P6del54R3 0.82% 2,981
EML4-ALK.E6aA20.AB374361 0.37% 2,492
PML-RARA.P6del12R3 0.01% 166
GOPC-ROS1.G4R36.COSF1188 0.02% 65
PRB212: EWSR1-FL1 %TMR Read Count
EWSR1-FLI1.E7F6 3.75% 20,815
EML4-ALK.E20A20.COSF409.1 1.11% 6,614
SLC34A2-ROS1.S4R32.COSF1196 0.29% 3,251
EML4-ALK.E6aA20.AB374361 0.38% 2,024
BCR-ABL1.B14A2 0.01% 58
PRB216: ETV6-NTRK3 %TMR Read Count
ETV6-NTRK3.E5N15.COSF571.1 3.43% 58,464
TCF3-PBX1.T16P3.COSF1489 0.06% 1,104
ETV6-NTRK3.E4N15.COSF823.2 0.00% 71
Representative bone marrow (starting with “CPM”) and FFPE (starting with “PRB”) samples with fusions
characterized by orthogonal methods. All expected primary isoforms detected.
Amplicons