Optimization of nested PCR conditions to improve target coverage uniformity in Archer™ VariantPlex™ BRCA Panel

AGBT Meeting, February 25-28, 2015

Authors

Laura Johnson and Brady Culver

ArcherDX, Inc., Boulder, CO


Abstract

Anchored Multiplex PCR (AMP™) chemistry for NGS applications is a robust tool to detect both known and unknown fusions. However, characteristics associated with AMP transcend RNA assays and make it a valuable tool for DNA sequencing as well. These characteristics include the following:

  • The ability to identify unique starting molecules and consolidate sequencing information from PCR duplicates
  • Low input DNA requirements
  • Applicability to both Illumina® and Ion Torrent™ sequencing platforms

Additionally, the enzymatic DNA fragmentation method used in Archer™ VariantPlex™ assays generates a broad sequenceable fragment size distribution ranging from less than 100bp to more than 1 kilobase in length. The presence of longer fragments facilitates read mapping over repetitive or conserved regions in the genome.

The first Archer DNA assay for widespread release is the Archer VariantPlex BRCA1/2 Panel. Inherited harmful BRCA1/2 mutations substantially increase the lifetime risk of developing cancer, especially breast and ovarian cancer. The VariantPlex BRCA1/2 Panel design covers all coding and non-coding exons in RefSeq annotated transcript models for these genes. In an effort to further improve panel performance, we investigated alterations to PCR cycling conditions that enhanced target coverage uniformity.

Our data shows that a balance must be struck between primer performance uniformity and the portion of total reads that are on-target. Despite losses in on-target reads, fewer reads were required to cover all targeted bases than conditions with higher on-target percent. In addition, coverage and sensitivity were universally improved by utilizing a new polymerase. These changes are being implemented across all Archer NGS kits. Finally, we demonstrate the performance of this panel and our variant calling software by correctly identifying the known BRCA1 or BRCA2 mutations present in 10 different samples from Coriell Institute.



Figure 1

BRCA Figure 1 Example coverage of BRCA1/2 Panel

Example coverage pattern of Archer VariantPlex BRCA1/2 Panel

The image shows example deduplicated read 1 coverage produced from the Archer VariantPlex BRCA1/2 Panel sequenced on an Illumina NextSeq® 500. The target area shown contains exons 10 and 11 of BRCA1, transcript variant NM_007294. Sequence reads were visualized using IGV (Robinson et al, Nature Biotechnology 2011). The VariantPlex BRCA1/2 panel was designed to cover all RefSeq-annotated coding and non-coding exons of BRCA1 and BRCA2. Additionally, exons unique to ENSEMBL predictions are covered in the panel. Example data was produced from 55ng single donor human adrenal genomic DNA (Biochain, Cat #D1234004) input and the Archer Universal DNA kit with VariantPlex BRCA1/2 primers. The enzymatic fragmentation step in DNA library construction produces sequence fragments ranging from 50bp to more than 1kb in length.


Figure 2

a) BRCA1/2 coverage depth improves with alterations to nested PCR cycling protocol.

The bar graph at right shows the percent of positions (y-axis) in RefSeq-annotated BRCA1 and BRCA2 variantscovered at the indicated depths (x-axis). Error bars show the standard deviation from three technical replicates. The data is taken from 250k paired-end reads. AMP utilizes a nested PCR approach to recapitulate the specificity achieved with conventional oposing primer-based PCR. Standard refers to our standard 20 cycles of the first PCR and 20 cycles of the second PCR; and modified 2 uses 12 cycles of the first PCR followed by 20 cycles of the second PCR. Library yeild was modestly reduced using fewer PCR 1 cycles (data not shown).


BRCA Figure 2A Coverage depth improves with alterations to nested PCR

b) Removal of PCR duplicates does not significantly reduce BRCA1/2 coverage depth.

The bar graph at left shows the percent of positions (y-axis) in RefSeq-annotated coding exons of BRCA1 and BRCA2 covered at the indicated depths (x-axis). Data was generated from 250k reads using the modified 2 cycling conditions. For all coverage calculations, coverage arising from primer sequence is not considered.

BRCA Figure 2B Removal of PCR duplicates does not reduce BRCA coverage depth

Figure 3

a) BRCA1/2 panel on-target percent is inversely related to PCR1 cycle number.

The bar graph shows the percent of reads in the BRCA1/2 panel that are on-target (y-axis) in relation to cycling condition tested (x-axis). On-target percentages are shown with (deduplicated reads) and without removal of PCR duplicates (total reads). Cycling conditions are the same as those described in Figure 2a. On-target is defined as a read that maps to its intended target location.

b) BRCA1/2 standard coverage uniformity metric improves with reduced PCR1 cycle number.

The bar graph shows the percent of positions (y-axis) in BRCA1 and BRCA2 that are at least 0.2x the mean coverage depth for the indicated cycling conditions (x-axis)

BRCA Figure3AB BRCA panel on-target percent is inversely related to PCR1 cycle number

c) BRCA1/2 coverage uniformity improves with reduced PCR1 cycle number.

The density plot shows the portion of positions (y-axis) in BRCA1 and BRCA2 with the indicated per-base ratio to median coverage for each cycling condition as indicated by colored line. A tight ditribution indicates a more uniform panel. The data depicts the total read coverage (no PCR duplicates removed).

BRCA Figure 3C BRCA coverage uniformity improves with reduces PCR1 cycle number

d) Removal of PCR duplicates produces a more uniform coverage distribution.

The density plot is as in 3c, but depicts uniformity based on the removal of pCR duplicates (blue) or not. Data is taken from 250k reads used modified 2.




BRCA Figure 3D Removal of PCR duplicates produces more uniform coverage distribution

Figure 4

DNA polymerase used for PCR steps affects library fragment size.

The histograms show library fragment size (x-axis) distributions (y-axis) for DNA polymerase A (left) and an alternative high-fidelity DNA ploymerase (polymerase B; right) found in libraries made using Archer VariantPlex BRCA1/2 panel.

BRCA Figure 4A Library Fragment Size
BRCA Figure 4B Library Fragment Size

Table

Archer Analysis software correctly calls 9/9 mutant samples from the Coriell Institute Biorepository. The table shows the sample information provided by Coriell and variant calling data produced by Archer Analysis using sequencing information derived from the indicated sample. SIFT and PolyPhen are two third-party algorithms designed to predict the consequences of mutations.

BRCA Table Archer Analysis correctly calls 9/9 mutant samples


Reference

James T. Robinson, Helga Thorvaldsdottir, Wendy Winckler, Mitchell Guttman, Eric S. Lander, Gad Getz, Jill P. Mesirov. Integrative Genomics Viewer. Nature Biotechnology. 29, 24-26 (2011)


For Research Use Only. NextSeq® and Illumina® are registered trademarks of Illumina, Inc. Ion Torrent™ is a trademark of Thermo Fisher Scientific. Archer™ and VariantPlex™ are trademarks of ArcherDX, Inc.

Subscribe to stay in the loop







How to contact us

Address

2477 55th Street, Suite 202

Boulder, CO 80301

Phone

Phone: (877) 771 1093

Phone: (303) 357 9001

All content © 2017 ArcherDX, Inc.

For Research Use Only. Not for use in diagnostic procedures. For Research Use Only. Not for use in diagnostic procedures.