Joshua Stahl

Detecting Gene Rearrangements in Patient Populations Through a 2-Step Diagnostic Test Comprised of Rapid IHC Enrichment Followed by Sensitive Next-Generation Sequencing.

Targeted therapy combined with companion diagnostics has led to the advancement of next-generation sequencing (NGS) for detection of molecular alterations. However, using a diagnostic test to identify patient populations with low prevalence molecular alterations, such as gene rearrangements, poses efficiency, and cost challenges. To address this, we have developed a 2-step diagnostic test to identify NTRK1, NTRK2, NTRK3, ROS1, and ALK rearrangements in formalin-fixed paraffin-embedded clinical specimens. This test is comprised of immunohistochemistry screening using a pan-receptor tyrosine kinase cocktail of antibodies to identify samples expressing TrkA (encoded by NTRK1), TrkB (encoded by NTRK2), TrkC (encoded by NTRK3), ROS1, and ALK followed by an RNA-based anchored multiplex polymerase chain reaction NGS assay. We demonstrate that the NGS assay is accurate and reproducible in identification of gene rearrangements. Furthermore, implementation of an RNA quality control metric to assess the presence of amplifiable nucleic acid input material enables a measure of confidence when an NGS result is negative for gene rearrangements. Finally, we demonstrate that performing a pan-receptor tyrosine kinase immunohistochemistry staining enriches detection of the patient population for gene rearrangements from 4% to 9% and has a 100% negative predictive value. Together, this 2-step assay is an efficient method for detection of gene rearrangements in both clinical testing and studies of archival formalin-fixed paraffin-embedded specimens.

SEC31A-ALK Fusion Gene in Lung Adenocarcinoma.

Anaplastic lymphoma kinase (ALK) fusion is a common mechanism underlying pathogenesis of non-small cell lung carcinoma (NSCLC) where these rearrangements represent important diagnostic and therapeutic targets. In this study, we found a new ALK fusion gene, SEC31A-ALK, in lung carcinoma from a 53-year-old Korean man. The conjoined region in the fusion transcript was generated by the fusion of SEC31A exon 21 and ALK exon 20 by genomic rearrangement, which contributed to generation of an intact, in-frame open reading frame. SEC31A-ALK encodes a predicted fusion protein of 1,438 amino acids comprising the WD40 domain of SEC31A at the N-terminus and ALK kinase domain at the C-terminus. Fluorescence in situ hybridization studies suggested that SEC31A-ALK was generated by an unbalanced genomic rearrangement associated with loss of the 3′-end of SEC31A. This is the first report of SEC31A-ALK fusion transcript in clinical NSCLC, which could be a novel diagnostic and therapeutic target for patients with NSCLC.

The Power Decoder Simulator for the Evaluation of Pooled shRNA Screen Performance

RNA interference screening using pooled, short hairpin RNA (shRNA) is a powerful, high-throughput tool for determining the biological relevance of genes for a phenotype. Assessing an shRNA pooled screen’s performance is difficult in practice; one can estimate the performance only by using reproducibility as a proxy for power or by employing a large number of validated positive and negative controls. Here, we develop an open-source software tool, the Power Decoder simulator, for generating shRNA pooled screening experiments in silico that can be used to estimate a screen’s statistical power. Using the negative binomial distribution, it models both the relative abundance of multiple shRNAs within a single screening replicate and the biological noise between replicates for each individual shRNA. We demonstrate that this simulator can successfully model the data from an actual laboratory experiment. We then use it to evaluate the effects of biological replicates and sequencing counts on the performance of a pooled screen, without the necessity of gathering additional data. The Power Decoder simulator is written in R and Python and is available for download under the GNU General Public License v3.0.

Abstract 3618: NGS-based detection of FLT3-ITDs with Anchored Multiplex PCR

FLT3 internal tandem duplications (ITDs) are found in > 20% of pediatric and adult acute myeloid leukemia (AML) cases and are generally associated with a poor prognosis. Nevertheless, detection of FLT3-ITDs presents a challenge to NGS-based approaches, as many variant callers fail to identify the highly variable repeated sequences associated with FLT3-ITDs. We have developed a novel targeted assay, the Archer™ VariantPlex™ Core AML Panel, that utilizes Anchored Multiplex PCR (AMP™) with probes in multiple locations proximal to exons 13, 14, and 15 of FLT3. These exons encompass the commonly mutated juxtamembrane domain and tyrosine kinase domain 1. This panel yields high-complexity, dual-strand coverage of known FLT3-ITD locations. Because AMP probes, unlike traditional PCR probes, function independently of each other, we are able to produce multiple overlapping “snapshots” of the region of interest, thereby enhancing the ability to confidently identify complex mutation types. The sequenced reads are assembled using a novel de novo assembly algorithm in the Archer Analysis bioinformatics pipeline, in which the resulting consensus is annotated and events that involve FLT3 exons 13 14 or 15 are marked as FLT3 abnormalities. In order to assess Archer Analysis and the VariantPlex Core AML panel, we tested it on >20 patient-derived samples with known FLT3-ITDs. In addition, we generated over two thousand in silico datasets, representing the spectrum of known ITDs, to further test our panel design and analysis algorithm. Each in silico dataset was constructed to simulate reads originating from probes in VariantPlex Core AML Panel. This methodology permitted rigorous optimization of both the VariantPlex Core AML Panel and the analysis algorithm. The VariantPlex Core AML Panel, in conjunction with our novel detection algorithm, showed both exceptional sensitivity and specificity in the detection of FLT3-ITDs. FLT3-ITDs were successfully identified in all patient samples tested, and no false positives were detected. Our in silico datasets showed similarly high sensitivity and specificity. These data indicate that AMP libraries targeting the FLT3-ITD region are ideal for the detection of this complex mutation type, largely because of the overlapping and anchored read structure. Therefore, the Archer VariantPlex Core AML panel, in conjunction with Archer Analysis represents a reliable platform for the detection of FLT3-ITDs, in addition to other mutations commonly found in AML. Citation Format: Marc Bessette, Benjamin Van Deusen, Michael Banos, Laura Johnson, Aaron Berlin, Erik Reckase, Joshua Stahl, Abel Licon, Brian Kudlow. NGS-based detection of FLT3-ITDs with Anchored Multiplex PCR. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3618.

Abstract 1381: NGS-based CNV detection sensitivity is dependent upon nucleic acid input quality

Copy number variations (CNV) impact more of the cancer genome than all other mutation types combined. Recent advances in next-generation sequencing (NGS) have enabled simultaneous detection of CNVs and other somatic mutations from FFPE-derived samples, but NGS-based detection of low level CNVs (ie 2-3x) remains challenging. Nucleic acid from FFPE is a common starting material for NGS-based cancer genotyping; however, this material is often of low complexity due to a variety of factors including limited mass amount, excessive fragmentation, or chemical crosslinking. Current practices often measure input mass, or the nanograms of DNA that are added to a reaction, yet it is input complexity, or the amount of nucleic acid available for NGS library generation, that truly dictates the amount of information that can be recovered from a given sample. Archer™ VariantPlex™ assays are targeted NGS panels that permit simultaneous detection of SNVs, in/dels, and CNVs using Anchored Multiplex PCR (AMP™). Molecular barcoded adapters are ligated to each input molecule prior to any amplification. This permits the unique identification of individual input molecules thus facilitating precise copy number measurements. In addition, AMP enables amplification of highly fragmented FFPE inputs as short fragments are captured between the ligated adapter and the enrichment probe. To determine the effect of input quality on sensitivity of CNV calling we characterized over 150 tumor sample input qualities and their resulting library metrics. In addition we modeled the effect of low tumor cellularity on CNV sensitivity by carrying out dilution experiments of CNV-positive samples into samples of normal copy number. Using Archer VariantPlex assays in conjunction with Archer Analysis, we have successfully detected CNVs as small as 2X in both FFPE and cell line DNA. We found that input nucleic acid quality, as measured by a qPCR-based assay called Archer PreSeq™ DNA QC, strongly impacted the sensitivity of CNV calling. Assessment of input complexity using the PreSeq DNA QC Assay is predictive of limit of detection for CNVs and identifies an input quantity that will result in high quality NGS libraries. Our dilution experiments confirmed the expected relationship between actual and measured copy number in our population-averaging assay. Nucleic acid damage typical of FFPE samples reduces CNV calling sensitivity; however, this loss of sensitivity can be partially mitigated by increasing the input quantity. This corroborates the notion that input complexity is the major driver of information-capture from NGS based assays. Finally, tumor cellularity displays a predictable effect on the measured CNV value. Citation Format: Josh Haimes, James Covino, Namitha Namoj, Elina Baravik, Laura Johnson, Joshua Stahl, Brady P. Culver, Brian Kudlow. NGS-based CNV detection sensitivity is dependent upon nucleic acid input quality. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1381.