Consensus Calling Research Articles

Primed and ready: nanopore metabarcoding can now recover highly accurate consensus barcodes that are generally indel-free

BackgroundDNA metabarcoding applies high-throughput sequencing approaches to generate numerous DNA barcodes from mixed sample pools for mass species identification and community characterisation. To date, however, most metabarcoding studies employ second-generation sequencing platforms like Illumina, which are limited by short read lengths and longer turnaround times. While third-generation platforms such as the MinION (Oxford Nanopore Technologies) can sequence longer reads and even in real-time, application of these platforms for metabarcoding has remained limited possibly due to the relatively high read error rates as well as the paucity of specialised software for processing such reads.ResultsWe show that this is no longer the case by performing nanopore-based, cytochrome c oxidase subunit I (COI) metabarcoding on 34 zooplankton bulk samples, and benchmarking the results against conventional Illumina MiSeq sequencing. Nanopore R10.3 sequencing chemistry and super accurate (SUP) basecalling model reduced raw read error rates to ~ 4%, and consensus calling with amplicon_sorter (without further error correction) generated metabarcodes that were ≤ 1% erroneous. Although Illumina recovered a higher number of molecular operational taxonomic units (MOTUs) than nanopore sequencing (589 vs. 471), we found no significant differences in the zooplankton communities inferred between the sequencing platforms. Importantly, 406 of 444 (91.4%) shared MOTUs between Illumina and nanopore were also found to be free of indel errors, and 85% of the zooplankton richness could be recovered after just 12–15 h of sequencing.ConclusionOur results demonstrate that nanopore sequencing can generate metabarcodes with Illumina-like accuracy, and we are the first study to show that nanopore metabarcodes are almost always indel-free. We also show that nanopore metabarcoding is viable for characterising species-rich communities rapidly, and that the same ecological conclusions can be obtained regardless of the sequencing platform used. Collectively, our study inspires confidence in nanopore sequencing and paves the way for greater utilisation of nanopore technology in various metabarcoding applications.

Using nanopore sequencing to identify fungi from clinical samples with high phylogenetic resolution

The study of microbiota has been revolutionized by the development of DNA metabarcoding. This sequence-based approach enables the direct detection of microorganisms without the need for culture and isolation, which significantly reduces analysis time and offers more comprehensive taxonomic profiles across broad phylogenetic lineages. While there has been an accumulating number of researches on bacteria, molecular phylogenetic analysis of fungi still remains challenging due to the lack of standardized tools and the incompleteness of reference databases limiting the accurate and precise identification of fungal taxa. Here, we present a DNA metabarcoding workflow for characterizing fungal microbiota with high taxonomic resolution. This method involves amplifying longer stretches of ribosomal RNA operons and sequencing them using nanopore long-read sequencing technology. The resulting reads were error-polished to generate consensus sequences with 99.5–100% accuracy, which were then aligned against reference genome assemblies. The efficacy of this method was explored using a polymicrobial mock community and patient-derived specimens, demonstrating the marked potential of long-read sequencing combined with consensus calling for accurate taxonomic classification. Our approach offers a powerful tool for the rapid identification of pathogenic fungi and has the promise to significantly improve our understanding of the role of fungi in health and disease.

ViralConsensus: A fast and memory-efficient tool for calling viral consensus genome sequences directly from read alignment data.

In viral molecular epidemiology, reconstruction of consensus genomes from sequence data is critical for tracking mutations and variants of concern. However, as the number of samples that are sequenced grows rapidly, compute resources needed to reconstruct consensus genomes can become prohibitively large. ViralConsensus is a fast and memory-efficient tool for calling viral consensus genome sequences directly from read alignment data. ViralConsensus is orders of magnitude faster and more memory-efficient than existing methods. Further, unlike existing methods, ViralConsensus can pipe data directly from a read mapper via standard input and performs viral consensus calling on-the-fly, making it an ideal tool for viral sequencing pipelines. ViralConsensus is freely available at https://github.com/niemasd/ViralConsensus as an open-source software project. Supplementary data are available at Bioinformatics online.

Abstract 3142: Enabling whole genome sequencing analysis from FFPE specimens in clinical oncology

Abstract Whole genome sequencing (WGS) enables the identification of all cancer associated biomarkers in a patient’s tumor genome. Whilst fresh frozen (FF) derived WGS data provides optimal data quality, the majority of clinical biospecimens are from formalin fixed paraffin embedded (FFPE) tissue which results in DNA damage and an increase in artifactual mutation calls. Development of analytical frameworks tailored to FFPE derived WGS data can unlock the potential of genome profiling in clinical oncology. We performed comprehensive WGS analysis on 58 matched FF/FFPE specimens derived from 3 cancer centers. Consensus calling detected high-confidence somatic mutations across variant classes including: single nucleotide variants (SNVs), insertions/deletions (indels), structural variants (SVs) and copy number aberrations (CNAs). For each sample, genome-wide mutational patterns including tumor mutational burden (TMB), SNV/indel signatures, and homologous recombination deficiency (HRD) scores were estimated. We developed a random forest based framework using 33 features to learn mutation patterns associated with FFPE artifacts and implemented a filtration strategy for FFPE derived WGS data within a clinical prototype analytical workflow. We identified a high degree of concordance (~91%, n=192/210) for oncogenic variants between FF/FFPE WGS data. Comparison of small mutation calls presented an average 2-fold increase in FFPE samples with a range up to 152x for SNVs and 43x for indels. However, this was not the case for SVs: -0.4x (range -0.8-1.4). We demonstrate that genome-wide mutation patterns were significantly affected, impacting estimates for TMB, HRD and signature contributions. On average, TMB was overestimated in FFPE (median=10.3, range: 1.4-94) versus FF (median=3.4, range:0.04-29.6). For 7 patients with evidence of HRD in FF, HRD scores did not reach statistical significance in FFPE. Mutational signatures in FFPE were enriched for COSMIC signatures 37 and 5. Our artifact classification model achieved ROC AUC of 97.5% and precision-recall of 98.9%. Post artifact filtration, precision in SNV/indel calling was increased from 49.3% to 93.4% and 61.8% to 82.7% respectively with no effect on driver alterations. This increased global signal concordance drastically, with comparable TMB scores (median 2.4; range .03-26.1) and improved cosine similarity for SNV/indel signatures (median 0.98; range 0.40-1). HRD was successfully detected in 7/7 patients from FFPE derived data post filtering with probability scores ranging from 0.76-1. We demonstrate that FFPE specimens harbor a variable increase in artifactual mutation burden in SNV/indels but not in SVs. We propose an effective filtering procedure which successfully removes FFPE related artifacts enabling accurate profiling of clinically relevant driver mutations and genome-wide mutation patterns from readily available FFPE-derived tumor specimens. Citation Format: Dylan Domenico, Gunes Gundem, Max F. Levine, Juanes E. Arango-Ossa, Pauline Robbe, Georgios Asimomitis, Cassidy Cobbs, Emily Stockfisch, Janine Senz, Dawn Cochrane, Neeman Mohibullah, Neerav Shukla, Sohrab P. Shah, Andrew McPherson, Anna Schuh, Andrew L. Kung, Elli Papaemmanuil. Enabling whole genome sequencing analysis from FFPE specimens in clinical oncology [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3142.

Arcch: Improving the Performance of Clonal Hematopoiesis Variant Calling and Interpretation Using a Consensus Based Approach

Introduction: The acquisition of somatic mutations in hematopoietic stem and progenitor stem cells with resultant clonal expansion is known as clonal hematopoiesis (CH). CH is linked to a higher risk of hematologic malignancies and other adverse health outcomes. The prevalence of CH is dependent on sequencing technique. With ultra-high-depth sequencing, CH can be detected in most adults. However, CH variant calling at very low variant allele fractions (VAF) is challenging due to difficulty in distinguishing low frequency CH mutations from sequencing artifacts. Here we develop and validate a novel CH variant calling approach that combines ensemble based variant calling with advanced artifact filtering and outperforms commonly used somatic variant callers. Methods: We sequenced a tumor-normal dilution series from six AML patients and 27 technical controls with a targeted panel utilizing unique molecular indexes that included nine common CH genes (DNMT3A, TET2, ASXL1, TP53, PPM1D, JAK2, SF3B1, SRSF2) with a mean unique coverage of 18,268x. In total we identified 32 mutations in these six samples validated by an independent sequencing panel performed for the purpose of clinical testing. The variant allele fraction (VAF) in the resulting tumor-dilution samples ranged from 0.1%-50% VAF. Results: For higher VAF mutations (>5%), Mutect2, Lofreq, and Vardict showed reasonable sensitivity (range: 0.70-1.00) and positive predictive values (PPV) (range: 0.80-1.00). However, at lower VAF, PPV was poor; for mutations between 0.1-0.4% VAF, the PPV ranged between 0.01-0.22. We applied additional filters based on sequencing quality, depth, regional complexity, and an empiric estimate of sequencing error at the position of a given called variant using technical controls. Taking the consensus of three callers and including our additional filters, we not only retained sensitivity and PPV for higher VAF variants above 1% (0.92 sensitivity, 0.95 PPV) but improved the PPV with retained sensitivity for variants between 1-0.1% VAF (0.91 sensitivity, 0.83 PPV). Applying our consensus approach to blood samples drawn from 31 healthy individuals sequenced using the same targeted panel with variants validated using an orthogonal approach, we saw excellent sensitivity and PPV across a wide range of VAFs (Sensitivity 0.99 [>1% VAF], 0.92 [0.2-1% VAF], and 0.82 [0.1-0.2% VAF]; PPV 1.00 [>1% VAF], 0.98 [0.2-1% VAF], 0.97 [0.2-0.1% VAF]). We compared our consensus calling and advanced filtering approach to a machine learning model. A decision-tree-based ensemble machine learning algorithm, XGBoost, was trained on the tumor-dilution samples with validation using 31 blood samples with orthogonal sequencing. The variant caller output (allele counts, flags, etc.), additional false positive filtering metrics, and statistical results from the technical controls were used as features. Sensitivity and PPV was high for mutations above 1% VAF (0.97 sensitivity and 0.82 PPV) with the performance declining for lower VAF mutations (0.72 sensitivity and 0.65 PPV). Thus, a machine learning approach did not outperform consensus calling with advanced artifact filtering. Finally, we developed a custom annotation pipeline for CH variant pathogenicity classification incorporating previously reported CH mutations in large cohorts, cancer somatic mutation frequency databases and curated cancer driver knowledge databases. Conclusion: We developed ArCCH, an advanced artifact filtering and consensus CH calling pipeline that includes custom annotation features to facilitate CH pathogenicity determination. Our results show that our consensus approach with advanced error correction filters substantially improves the performance of CH variant calling beyond commonly used single variant callers. This end-to-end flexible workflow package is publicly available on Terra and will facilitate future CH studies.

Variability in Pattern of Mutational Signatures in Multiple Myeloma As a Function of Racial Origin

Introduction: Current data suggests Black/African Americans (AA) have an increased risk of developing multiple myeloma (MM), especially in comparison to White/European Americans (EA). However, the underlying causes remain unclear. Genetic definition of the spectrum of acquired mutations and the signatures they contain could help us to understand the mechanism underlying this excess risk. Currently, only limited information is available on the mutational spectra within AAs and the majority of data that has been generated is based on the societal construct of self-identified race/ethnicity alone. To address this deficiency and the potential associated bias, it is critical to examine greater numbers of whole genome sequencing (WGS) data from AAs. Data from a pilot study indicate a prolonged and more intense exposure to the germinal center (GC) reaction in the early evolutionary phases of MM within AAs, suggesting that it is essential to examine sequencing data from early disease stage i.e. smoldering MM (SMM) to better understand genetic predisposition in AAs. Methods: To address the mutational basis of MM and how it varies based on racial origin we carried out two large studies, Polyethnic-1000 and the Smoldering Myeloma Registration Trial (SMRT) study. Furthermore, we examined a cross-sectional series of WGS and whole exome sequencing (WES) data derived from cases with SMM and MM already sequenced to characterize the pattern of mutations and the signatures in both the coding and non-coding regions of the genome. Importantly, to accurately infer racial ancestry, we built an admixture workflow into our sequencing analysis pipeline, enabling us to address the variability in origin from within the MM cases ranging from regions of Europe to subsets of cases with African, North/South American, and Asian origin directly. We showed its use in previous work describing the heterogenous genetic admixture in MM cases that self-identified Hispanic and Latino. Our pipeline uses a suite of well-supported bioinformatic tools for preprocessing sequencing data according to best practices, cataloguing germline variants across a large series of patients, generating ancestry admixture estimations per patient, determining telomere length and composition, and identifying the spectrum of somatic events including single nucleotide variants, small insertions/deletions, copy number variants, and structural variants using consensus calling. All source code and reference data is publicly available via a GitHub repository (https://github.com/pblaney/mgp1000). Results: We generated new WGS data from 65 MM cases (EA n=15, AA n=50), and 44 SMM cases (EA n=36, AA n=18) as part of the Polyethnic 1000 and SMRT study. We also systematically reanalyzed 473 cases with deep (>30x) WGS data and all cases in the CoMMpass study low-depth (5-12x) WGS (n=943). Estimating ancestral composition directly from the genetic data of the CoMMpass study, we identified discrepancies between the self-reported race and the genetically determined admixture composition. Approximately 75% of individuals identified as 'White' but 7% of these cases had a genetic composition that included greater than 10% contribution from African, North/South American, or Asian ancestry. More interesting, 18% identified as 'Black/African American' but 50% of these cases had greater than 10% contribution from European, North/South American, or Asian ancestry. These results demonstrate the underlying genetic heterogeneity in a population that is currently clinically classified as the homogenous and thus lead to skewed evaluations of race and outcome in MM. We will compare the mutational spectra within these groups with different origins to address the risk of developing MM as well as addressing the time of origin of MM within the different populations. Conclusion: We show that there is significant variability of admixture from Europe, North/South America, and Africa within patients who self-identify as White, Black/African American, or Hispanic/Latino in clinical datasets from the U.S. This reinforces the need for complementary genetic admixture with racial self-identification when investigating the links between racial ancestry and outcomes in MM. Figure: Admixture composition of CoMMpass study shows significant contribution of African, North/South American, and East/South Asian origins in predominant European ancestral background. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Accurate gene consensus at low nanopore coverage.

Nanopore technologies allow high-throughput sequencing of long strands of DNA at the cost of a relatively large error rate. This limits its use in the reading of amplicon libraries in which there are only a few mutations per variant and therefore they are easily confused with the sequencing noise. Consensus calling strategies reduce the error but sacrifice part of the throughput on reading typically 30 to 100 times each member of the library. In this work, we introduce SINGLe (SNPs In Nanopore reads of Gene Libraries), an error correction method to reduce the noise in nanopore reads of amplicons containing point variations. SINGLe exploits that in an amplicon library, all reads are very similar to a wild-type sequence from which it is possible to experimentally characterize the position-specific systematic sequencing error pattern. Then, it uses this information to reweight the confidence given to nucleotides that do not match the wild-type in individual variant reads and incorporates it on the consensus calculation. We tested SINGLe in a mutagenic library of the KlenTaq polymerase gene, where the true mutation rate was below the sequencing noise. We observed that contrary to other methods, SINGLe compensates for the systematic errors made by the basecallers. Consequently, SINGLe converges to the true sequence using as little as 5 reads per variant, fewer than the other available methods.

The genetic mechanisms underlying the convergent evolution of pollination syndromes in the Neotropical radiation of Costus L.

Selection together with variation in floral traits can act to mold floral form, often driven by a plant’s predominant or most effective pollinators. To investigate the evolution of traits associated with pollination, we developed a phylogenetic framework for evaluating tempo and mode of pollination shifts across the genus Costus L., known for its evolutionary toggle between traits related to bee and bird pollination. Using a target enrichment approach, we obtained 957 loci for 171 accessions to expand the phylogenetic sampling of Neotropical Costus. In addition, we performed whole genome resequencing for a subset of 20 closely related species with contrasting pollination syndromes. For each of these 20 genomes, a high-quality assembled transcriptome was used as reference for consensus calling of candidate loci hypothesized to be associated with pollination-related traits of interest. To test for the role these candidate genes may play in evolutionary shifts in pollinators, signatures of selection were estimated as dN/dS across the identified candidate loci. We obtained a well-resolved phylogeny for Neotropical Costus despite conflict among gene trees that provide evidence of incomplete lineage sorting and/or reticulation. The overall topology and the network of genome-wide single nucleotide polymorphisms (SNPs) indicate that multiple shifts in pollination strategy have occurred across Costus, while also suggesting the presence of previously undetected signatures of hybridization between distantly related taxa. Traits related to pollination syndromes are strongly correlated and have been gained and lost in concert several times throughout the evolution of the genus. The presence of bract appendages is correlated with two traits associated with defenses against herbivory. Although labellum shape is strongly correlated with overall pollination syndrome, we found no significant impact of labellum shape on diversification rates. Evidence suggests an interplay of pollination success with other selective pressures shaping the evolution of the Costus inflorescence. Although most of the loci used for phylogenetic inference appear to be under purifying selection, many candidate genes associated with functional traits show evidence of being under positive selection. Together these results indicate an interplay of phylogenetic history with adaptive evolution leading to the diversification of pollination-associated traits in Neotropical Costus.

Mitochondrial DNA Consensus Calling and Quality Filtering for Constructing Ancient Human Mitogenomes: Comparison of Two Widely Applied Methods.

Retrieving high-quality endogenous ancient DNA (aDNA) poses several challenges, including low molecular copy number, high rates of fragmentation, damage at read termini, and potential presence of exogenous contaminant DNA. All these factors complicate a reliable reconstruction of consensus aDNA sequences in reads from high-throughput sequencing platforms. Here, we report findings from a thorough evaluation of two alternative tools (ANGSD and schmutzi) aimed at overcoming these issues and constructing high-quality ancient mitogenomes. Raw genomic data (BAM/FASTQ) from a total of 17 previously published whole ancient human genomes ranging from the 14th to the 7th millennium BCE were retrieved and mitochondrial consensus sequences were reconstructed using different quality filters, with their accuracy measured and compared. Moreover, the influence of different sequence parameters (number of reads, sequenced bases, mean coverage, and rate of deamination and contamination) as predictors of derived sequence quality was evaluated. Complete mitogenomes were successfully reconstructed for all ancient samples, and for the majority of them, filtering substantially improved mtDNA consensus calling and haplogroup prediction. Overall, the schmutzi pipeline, which estimates and takes into consideration exogenous contamination, appeared to have the edge over the much faster and user-friendly alternative method (ANGSD) in moderate to high coverage samples (>1,000,000 reads). ANGSD, however, through its read termini trimming filter, showed better capabilities in calling the consensus sequence from low-quality samples. Among all the predictors of overall sample quality examined, the strongest correlation was found for the available number of sequence reads and bases. In the process, we report a previously unassigned haplogroup (U3b) for an Early Chalcolithic individual from Southern Anatolia/Northern Levant.

EDGE COVID-19: a web platform to generate submission-ready genomes from SARS-CoV-2 sequencing efforts.

SummaryGenomics has become an essential technology for surveilling emerging infectious disease outbreaks. A range of technologies and strategies for pathogen genome enrichment and sequencing are being used by laboratories worldwide, together with different and sometimes ad hoc, analytical procedures for generating genome sequences. A fully integrated analytical process for raw sequence to consensus genome determination, suited to outbreaks such as the ongoing COVID-19 pandemic, is critical to provide a solid genomic basis for epidemiological analyses and well-informed decision making. We have developed a web-based platform and integrated bioinformatic workflows that help to provide consistent high-quality analysis of SARS-CoV-2 sequencing data generated with either the Illumina or Oxford Nanopore Technologies (ONT). Using an intuitive web-based interface, this workflow automates data quality control, SARS-CoV-2 reference-based genome variant and consensus calling, lineage determination and provides the ability to submit the consensus sequence and necessary metadata to GenBank, GISAID and INSDC raw data repositories. We tested workflow usability using real world data and validated the accuracy of variant and lineage analysis using several test datasets, and further performed detailed comparisons with results from the COVID-19 Galaxy Project workflow. Our analyses indicate that EC-19 workflows generate high-quality SARS-CoV-2 genomes. Finally, we share a perspective on patterns and impact observed with Illumina versus ONT technologies on workflow congruence and differences.Availability and implementation https://edge-covid19.edgebioinformatics.org, and https://github.com/LANL-Bioinformatics/EDGE/tree/SARS-CoV2.Supplementary information Supplementary data are available at Bioinformatics online.

PacRAT: a program to improve barcode-variant mapping from PacBio long reads using multiple sequence alignment.

Use of PacBio sequencing for characterizing barcoded libraries of genetic variants is on the rise. However, current approaches in resolving PacBio sequencing artifacts can result in a high number of incorrectly identified or unusable reads. Here, we developed a PacBio Read Alignment Tool (PacRAT) that improves the accuracy of barcode-variant mapping through several steps of read alignment and consensus calling. To quantify the performance of our approach, we simulated PacBio reads from eight variant libraries of various lengths and showed that PacRAT improves the accuracy in pairing barcodes and variants across these libraries. Analysis of real (non-simulated) libraries also showed an increase in the number of reads that can be used for downstream analyses when using PacRAT. PacRAT is written in Python and is freely available (https://github.com/dunhamlab/PacRAT). Supplemental data are available at Bioinformatics online.

Accurate detection of circulating tumor DNA using nanopore consensus sequencing

Levels of circulating tumor DNA (ctDNA) in liquid biopsies may serve as a sensitive biomarker for real-time, minimally-invasive tumor diagnostics and monitoring. However, detecting ctDNA is challenging, as much fewer than 5% of the cell-free DNA in the blood typically originates from the tumor. To detect lowly abundant ctDNA molecules based on somatic variants, extremely sensitive sequencing methods are required. Here, we describe a new technique, CyclomicsSeq, which is based on Oxford Nanopore sequencing of concatenated copies of a single DNA molecule. Consensus calling of the DNA copies increased the base-calling accuracy ~60×, enabling accurate detection of TP53 mutations at frequencies down to 0.02%. We demonstrate that a TP53-specific CyclomicsSeq assay can be successfully used to monitor tumor burden during treatment for head-and-neck cancer patients. CyclomicsSeq can be applied to any genomic locus and offers an accurate diagnostic liquid biopsy approach that can be implemented in clinical workflows.

Genomic and Molecular Analyses Identify Molecular Subtypes of Pancreatic Cancer Recurrence

Genomic and Molecular Analyses Identify Molecular Subtypes of Pancreatic Cancer Recurrence

Ryūtō: improved multi-sample transcript assembly for differential transcript expression analysis and more.

Accurate assembly of RNA-seq is a crucial step in many analytic tasks such as gene annotation or expression studies. Despite ongoing research, progress on traditional single sample assembly has brought no major breakthrough. Multi-sample RNA-Seq experiments provide more information than single sample datasets and thus constitute a promising area of research. Yet, this advantage is challenging to utilize due to the large amount of accumulating errors. We present an extension to Ryūtō enabling the reconstruction of consensus transcriptomes from multiple RNA-seq datasets, incorporating consensus calling at low level features. We report stable improvements already at three replicates. Ryūtō outperforms competing approaches, providing a better and user-adjustable sensitivity-precision trade-off. Ryūtō's unique ability to utilize a (incomplete) reference for multi sample assemblies greatly increases precision. We demonstrate benefits for differential expression analysis.Ryūtō consistently improves assembly on replicates of the same tissue independent of filter settings, even when mixing conditions or time series. Consensus voting in Ryūtō is especially effective at high precision assembly, while Ryūtō's conventional mode can reach higher recall. Ryūtō is available at https://github.com/studla/RYUTO. Supplementary data are available at Bioinformatics online.

AbPOA: an SIMD-based C library for fast partial order alignment using adaptive band.

Partial order alignment, which aligns a sequence to a directed acyclic graph, is now frequently used as a key component in long-read error correction and assembly. We present abPOA (adaptive banded Partial Order Alignment), a Single Instruction Multiple Data (SIMD)-based C library for fast partial order alignment using adaptive banded dynamic programming. It can work as a stand-alone multiple sequence alignment and consensus calling tool or be easily integrated into any long-read error correction and assembly workflow. Compared to a state-of-the-art tool (SPOA), abPOA is up to 10 times faster with a comparable alignment accuracy. abPOA is implemented in C. A stand-alone tool and a C/Python software interface are freely available at https://github.com/yangao07/abPOA. Supplementary data are available at Bioinformatics online.

Deterrence in the age of artificial intelligence & autonomy: a paradigm shift in nuclear deterrence theory and practice?

ABSTRACT How might nuclear deterrence be affected by the proliferation of artificial intelligence (AI) and autonomous systems? How might the introduction of intelligent machines affect human-to-human (and human-to-machine) deterrence? Are existing theories of deterrence still applicable in the age of AI and autonomy? The article builds on the rich body of work on nuclear deterrence theory and practice and highlights some of the variegated and contradictory – especially human cognitive psychological – effects of AI and autonomy for nuclear deterrence. It argues that existing theories of deterrence are not applicable in the age of AI and autonomy and introducing intelligent machines into the nuclear enterprise will affect nuclear deterrence in unexpected ways with fundamentally destabilising outcomes. The article speaks to a growing consensus calling for conceptual innovation and novel approaches to nuclear deterrence, building on nascent post-classical deterrence theorising that considers the implications of introducing non-human agents into human strategic interactions.

Abstract 1988: Enabling personalized biomarker discovery in challenging oncology samples by coupling a novel library preparation chemistry with hybridization capture

Abstract Discovery and identification of new, targetable biomarkers is driven by comprehensive tumor profiling using next generation sequencing (NGS). However, converting tissue samples into NGS libraries is often challenging due to the low quantity and quality of DNA in such samples. Here we present sensitive and accurate detection of low-frequency variants by combining a novel library preparation optimized for low-input and degraded samples with IDT's xGen hybridization capture. The library preparation is enabled by an engineered mutant ligase and proprietary sequencing adapters that together prevent chimeras, suppress dimer-formation, enable double strand consensus calling, and maximize conversion. We demonstrate the performance of these reagents with archived tissue samples collected from five individuals with lung cancer. These matched samples consisted of DNA isolated from formalin-fixed paraffin-embedded (FFPE) tumor, DNA isolated from fresh-frozen normal tissue, and cell-free DNA (cfDNA) isolated from plasma. Sequencing libraries were prepared from 250 ng of tumor and healthy samples and 10 ng of cfDNA. Libraries were captured using a pan-cancer panel designed to simultaneously detect copy number variations, indels, rearrangements, and microsatellite instability across 532 oncogene targets. For each individual, tumor-associated mutations were identified using single-stranded consensus calling of variants found in tumor but not normal samples. As cfDNA has been correlated to disease progression, we wanted to determine if tumor-associated variants could be identified in the matched cfDNA samples. Custom capture panels were designed and delivered within 5 business days to target each individual's tumor-associated mutations. Relevant cfDNA samples were captured with these custom panels, sequenced, and variants were called after double stranded consensus calling. Mutations called from both FFPE tumor and cfDNA samples were verified by droplet digital PCR. By combining a high conversion library preparation with efficient hybridization capture and double stranded consensus calling, we have demonstrated an effective approach to extract information from difficult samples. Citation Format: Karissa Scott, Ushati Das Chakravarty, Hsiao-Yun Huang, Timothy Barnes, Kevin Lai, Jessica Sheu, Tzu-Chun Chen, Ramses Lopez, Lynette Lewis, Anastasia Potts, Steven Henck. Enabling personalized biomarker discovery in challenging oncology samples by coupling a novel library preparation chemistry with hybridization capture [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1988.

Evaluation of a regularly updated knowledge base for curation of somatic mutations detected in whole exomes of melanoma and lung, colorectal, and breast cancers.

e14072 Background: Evolving medical guidelines and complex multi-variant data from next-generation sequencing (NGS) testing of cancer samples make routine clinical interpretation of somatic variants challenging. We assessed the ability of NAVIFY(R) Mutation Profiler*, a CE-IVD somatic variant interpretation tool, to yield accurate time- and geography-specific clinical content on 2511 samples from The Cancer Genome Atlas (TCGA) across six solid tumor types. Methods: Whole exomes from lung adenocarcinoma (n = 469), lung squamous cell carcinoma (n = 325), colon adenocarcinoma (n = 368), rectum adenocarcinoma (n = 149), breast invasive carcinoma (n = 806), and skin cutaneous melanoma (n = 394) cases were analyzed. We utilized TCGA data from the Multi-Center Mutation Calling in Multiple Cancers (MC3) project to obtain consensus calling results of single nucleotide variants and indels. The open-access Mutation Annotation Format (MAF) file (v0.2.8) that stores variant calls was lifted to human reference genome GRCh38 and converted to individual Variant Call Format (VCF) files per case. VCF files were uploaded to NAVIFY Mutation Profiler to interpret actionable mutations according to a highly curated and up-to-date knowledge base (Roche Content v2.13.0 released December 6, 2019). We further assessed the accuracy of interpreting co-occurrences of actionable mutations. Results: Over 1.24 million somatic mutations across 20,590 genes were assessed with NAVIFY Mutation Profiler, which reported tier classifications of variants based on consensus recommendations from AMP, ASCO, CAP, and ACMG. 86% of cases had variants of strong (Tier I-A or I-B) or potential (Tier II-C or II-D) clinical significance; 56% of these cases had Tier I classifications, supported by robust clinical evidence. Potentially actionable variant-variant interactions were found in 14% of cases. The tool also identified appropriate tier classifications by geographic region in accordance with local medical guidelines. Conclusions: To benchmark against other tools, we utilized available exome data from TCGA MC3 to assess NAVIFY Mutation Profiler. While this study likely underestimates the fraction of cases with actionable mutations, given that copy number alterations or rearrangements are also present in TCGA samples, we found a higher yield of potentially actionable annotation than other published methods. * This product has not been evaluated by the Food and Drug Administration and is not commercially available in the United States.

Detection of Oncogenic and Clinically Actionable Mutations in Cancer Genomes Critically Depends on Variant Calling Tools

The analysis of cancer genomes provides fundamental information about its aetiology, the processes driving cell transformation or potential treatments. While researchers and clinicians are often only interested in the identification of oncogenic mutations, actionable variants or mutational signatures, the first crucial step in the analysis of any tumor genome is the identification of somatic variants in cancer cells (i.e. those that have been acquired during their evolution). For that purpose, a wide range of computational tools have been developed in recent years to detect somatic mutations in sequencing data from tumors. While there have been some efforts to benchmark somatic variant calling tools and strategies, the extent to which variant calling decisions impact the results of downstream analyses of tumor genomes remains unknown. Here we present a study to elucidate whether different variant callers (MuSE, MuTect2, SomaticSniper, VarScan2) and strategies to combine them (Consensus and Union) lead to different results in three important analyses of cancer genomics data: identification of cancer driver genes, quantification of mutational signatures and detection of clinically actionable variants. To this end, we tested how the results of these three analyses varied depending on the somatic mutation caller or strategy in five different projects from The Cancer Genome Atlas (TCGA). Our results show that variant calling decisions have a significant impact on these analyses, creating important differences that could even impact treatment decisions for some patients. Moreover, the Consensus calling strategy to combine the output of multiple variant calling tools, a very widely used strategy by the research community, can lead to the loss of some cancer driver genes and actionable mutations. On the other hand, the Union seems to be a legit strategy for some downstream analyses with a robust performance. Overall, our results point to important differences in critical analyses of tumor sequencing data depending on variant calling and highlight the limitations of widespread practices within the cancer genomics community.

TideHunter: efficient and sensitive tandem repeat detection from noisy long-reads using seed-and-chain.

MotivationPacific Biosciences (PacBio) and Oxford Nanopore Technologies (ONT) sequencing technologies can produce long-reads up to tens of kilobases, but with high error rates. In order to reduce sequencing error, Rolling Circle Amplification (RCA) has been used to improve library preparation by amplifying circularized template molecules. Linear products of the RCA contain multiple tandem copies of the template molecule. By integrating additional in silico processing steps, these tandem sequences can be collapsed into a consensus sequence with a higher accuracy than the original raw reads. Existing pipelines using alignment-based methods to discover the tandem repeat patterns from the long-reads are either inefficient or lack sensitivity.ResultsWe present a novel tandem repeat detection and consensus calling tool, TideHunter, to efficiently discover tandem repeat patterns and generate high-quality consensus sequences from amplified tandemly repeated long-read sequencing data. TideHunter works with noisy long-reads (PacBio and ONT) at error rates of up to 20% and does not have any limitation of the maximal repeat pattern size. We benchmarked TideHunter using simulated and real datasets with varying error rates and repeat pattern sizes. TideHunter is tens of times faster than state-of-the-art methods and has a higher sensitivity and accuracy.Availability and implementationTideHunter is written in C, it is open source and is available at https://github.com/yangao07/TideHunter