1. The Right Tool for the Job: Two Platforms for Targeted DNA Sequencing
AGBT 2018: #811
The Right Tool for the Job: Two Platforms for Targeted DNA Sequencing
Cassie Schumacher, Ashley Wood, Sukhinder Sandhu, Justin Lenhart, Laurie Kurihara, Vladimir Makarov
Swift Biosciences Inc, 58 Parkland Plaza, Suite 100, Ann Arbor, MI 48103
Abstract
Panel Specifications
Next-generation sequencing of pharmacogenetic and cancer-associated genes is valuable to advance our understanding of disease progression and treatment. However, obstacles including homology to pseudogenes and difficult to sequence motifs arise, making genes like CYP2D6, BRCA1, and BRCA2 challenging to sequence with short-read (SR) chemistry. Long-read (LR) assays overcome these issues, but are restricted due to limited available technologies. Both techniques are suitable for high-throughput sample processing. To enhance the benefits of each platform, we developed targeted, single-tube, multiplexed amplicon assays that allow for overlapping, contiguous coverage.
DNA (Coriell Institute) with known CYP2D6 variants was used in the 364-amplicon (~40kB) SR assay, targeting a variety of pharmacogenetic variants, and LR assay (~6.5kB) covering 100% of the CYP2D6 gene and upstream regulatory sequence. SR and LR libraries were sequenced on an Illumina MiniSeq® and PacBio RSII, respectively. Libraries from both assays had >90% on-target and coverage uniformity. 92% of known CYP2D6 variants were detected using the SR assay while all variants were detected with the LR technique.
DNA from BC01 (Coriell Institute) was used in the 246-amplicon (~23Kb) SR assay and the 35-amplicon LR assay (~86kB). While both comprehensively cover all exons, the LR assay additionally covers 20 complete introns. Greater than 95% on-target and coverage uniformity was seen in libraries from both SR and LR assays, and the same number of variants were detected.
Use of the correct tool is essential to evaluate genetic variation in disease and drug response. SR assays allow for screening of a wide range of targets for immediate and broad use. LR assays enable comprehensive sequencing of entire gene coding regions and neighboring introns, gaining additional insight into variants in difficult-to-sequence regions, large insertion/deletions, and structural variations. A
Short-read B
Short-read
Long-read C D
Panel
# of Amplicons
Amplicon Size (kb)
Panel Size (kb)
Tiling?
ADME - Short
364
0.15 40
Yes
CYP2D6- Long 1
6.5 No
BRCA1/2 - Short
246 23
BRCA1/2 - Long 36
2.5 90
Figure 3. Integrative Genome Viewer (IGV, Broad Institute) genome-wide view of the 364 pharmacogenetic short-read panel targets (A) IGV view of short- and long-read panel target coverage of the CYP2D6 (B), BRCA1 (C), and BRCA2 (D) genes. Short- and long-read panel specifications (E). Note that the entire exonic regions of both BRCA1 and BRCA2 are covered by both panels, while the entire exonic region of CYP2D6 is covered only by the long-read panel due to high pseudogene homology.
Sequencing Metrics
CYP2D6
# Aligned Reads
% On Target
Coverage Uniformity
DNA
Short
Long
NA17084
715
94.1
99.8
91.5
100
NA17221
535
94.3
99.7
NA17205
632
93.6
92.2
NA17293
507
91.9
NA17230
631
93.3
92.3
NA12244
538
93.9
92.0
NA17272
663
93.0
91.8
NA17039
334
93.1
NA17269
739
92.9
NA17276
254
NA17281
285
93.8
91.2
NA17204
593
BRCA1/2
# Aligned Reads
% On Target
Coverage Uniformity
DNA
Short
Long
NA14623
6203
99%
98%
100%
NA14624
6061
NA14626
6497
97%
NA13705
2416
NA13715
2070
NA14090
2449
96%
NA14094
2356
NA14638
2526
86%
NA14634
2686
NA14636
3128
NA14637
3010
NA14170
3381
Figure 4.Tables list the number of aligned reads, percent of reads on target, and the coverage uniformity in both assay types. All DNA samples were obtained from the Coriell Institute. All short-read samples were sequenced on an Illumina MiniSeq. CYP2D6 long read samples were sequenced on a PacBio RS II, while BRCA1/2 long-read samples were sequenced on a PacBio Sequel™. Aligned reads for all long-read samples are derived from CCS read numbers.
Dual Platform Approach
Short-Read Workflow
Long-Read Workflow
PCR 1
PCR 2
Ligation
Variant Analysis Using Short- and Long-Read Technology
CYP2D6
DNA
Haplotype
Variant
Expected AF
Observed AF
Short-Read
Long-Read
NA17084
*1/*10
P34S
0.5
0.68
0.51
S486T
0.52
0.57
R296C
0.69
0.47
NA17221
*1XN/*2
0.65
0.34
0.61
0.39
NA17205
*1/*41
0.45
0.53
0.49
0.55
NA17293
*2/*9
K281del
0.56
NA17230
*4/*41
0.54
L91M
0.48
H94R
0.23
0.4 1
NA12244
*35/*41
V11M
0.50
NA17272
*4/*10
1.00
0.96
NA17039
*2/*17
0.99
T107I
NA17269
*2/*41
NA17276
*2/*5
NA17281
*5/*9
0.98
NA17204
*1/*35
0.46
BRCA1/2
DNA
Gene
Variant
Expected AF
Observed AF
Short-Read
Long-Read
NA14623
BRCA2
TYR42CYS
0.50
0.53
0.47
NA14624
5946delCT
0.52
0.49
NA14626
LYS3326TER
0.54
NA13705
BRCA1
4-BP DEL, FS1252TER
0.31
0.45
NA13715
1-BP INS, 5382C
0.43
NA14090
2-BP DEL, 185AG
0.46
NA14094
40-BP DEL, FS397TER
0.44
0.56
NA14638
IVS5-11T>G
NA14634
4-BP DEL, FS1364TER
0.51
NA14636
5677insA
NA14637
ARG1443TER
0.48
NA14170
1-BP DEL, 6174T, FS
0.55
Figure 1. The short-read workflow consists of a multiplexed PCR followed by an indexing step to barcode and adapt the PCR products for Illumina sequencing (Left). The long-read workflow includes two PCR steps to amplify the target sequences and a ligation step to barcode and adapt the PCR products for PacBio sequencing. (Right)
Figure 5. Known variants for all 12 samples in both CYP2D6 BRCA1/2 genes are detailed. The expected allele frequency (AF) is given, along with the observed AF in both the short-read and long-read assay. All known variants were identified using both techniques.
Sequencing Challenges with Pseudogenes
Conclusion
Accel-Amplicon™ panels from Swift Biosciences can be used on short- and long-read sequencing technologies for variant detection.
Short-read amplicon panels are useful tools to interrogate variants of known significance in coding regions and intron/exon boundaries.
Long-read amplicon panels are useful for full gene coverage to not only analyze variants in the coding region, but to also probe neighboring introns which can be difficult to target with short-read amplicons due to repetitive regions and low complexity motifs.
Long-read technology provides more accurate alignment to identify structural variants and can overcome pseudogene alignment artifacts.
Long-read technology enables phasing of the CYP2D6 gene within the 6.5 kb single amplicon.
Both short- and long-read amplicon panel workflows can be completed in a single day and have the ability to process samples in a high-throughput manner.
Figure 2. Short-read sequencing poses a significant challenge when it comes to genes with known pseudogenes. Even when care has been taken to design primers to specific, unique regions, reads can still align to the pseudogene. Long-read technology can include much more unique content and additional bases, therefore all of the reads align on target. The figures show short-read and long-read targets and aligned reads to CYP2D6 (top) and the CYP2D7 pseudogene (bottom) for NA02016.
Targets
Reads
Short
Long
Targets
Reads
Short
Long
Swift Biosciences, Inc.
58 Parkland Plaza, Suite 100 • Ann Arbor, MI wwwswiftbiosci.com
© 2018, Swift Biosciences, Inc. The Swift logo and Accel-Amplicon are trademarks and Accel-NGS is a registered trademark of Swift Biosciences. This product is for Research Use Only. Not for use in diagnostic procedures. Illumina and MiniSeq are registered trademarks of Illumina, Inc. Sequel is a trademark of Pacific Biosciences , 02/18