Haplotype Caller Research Articles

Exploring the Role of Non-synonymous and Deleterious Variants Identified in Colorectal Cancer: A Multi-dimensional Computational Scrutiny of Exomes.

Colorectal cancers are the world's third most commonly diagnosed type of cancer. Currently, there are several diagnostic and treatment options to combat it. However, a delay in detection of the disease is life-threatening. Additionally, a thorough analysis of the exomes of cancers reveals potential variation data that can be used for early disease prognosis. By utilizing a comprehensive computational investigation, the present study aimed to reveal mutations that could potentially predispose to colorectal cancer. Ten colorectal cancer exomes were retrieved. Quality control assessments were performed using FastQC and MultiQC, gapped alignment to the human reference genome (hg19) using Bowtie2 and calling the germline variants using Haplotype caller in the GATK pipeline. The variants were filtered and annotated using SIFT and PolyPhen2 successfully categorized the mutations into synonymous, non-synonymous, start loss and stop gain mutations as well as marked them as possibly damaging, probably damaging and benign. This mutational profile helped in shortlisting frequently occurring mutations and associated genes, for which the downstream multi-dimensional expression analyses were carried out. Our work involved prioritizing the non-synonymous, deleterious SNPs since these polymorphisms bring about a functional alteration to the phenotype. The top variations associated with their genes with the highest frequency of occurrence included LGALS8, CTSB, RAD17, CPNE1, OPRM1, SEMA4D, MUC4, PDE4DIP, ELN and ADRA1A. An in-depth multi-dimensional downstream analysis of all these genes in terms of gene expression profiling and analysis and differential gene expression with regard to various cancer types revealed CTSB and CPNE1 as highly expressed and overregulated genes in colorectal cancer. Our work provides insights into the various alterations that might possibly lead to colorectal cancer and suggests the possibility of utilizing the most important genes identified for wet-lab experimentation.

Application of a custom haplotype caller to analyze sequence-based data of 56 microhaplotypes

Microhaplotypes (microhaps) are recently introduced markers that aim to complement the limitations of conventional forensic markers such as short tandem repeats (STRs) and single nucleotide polymorphisms (SNPs). With the potential of microhaps in forensics becoming clearer through massively parallel sequencing (MPS), MPS-based studies on microhaps are being actively reported. However, simpler workflow schemes for the generation and analysis of MPS data are still required to facilitate the practical application of MPS in forensics. In this study, we developed an in-house MPS panel that simultaneously amplifies 56 microhaps and a custom haplotype caller, Visual Microhap. The developed tool works on a web browser and provides four analysis options to extract SNP-based haplotypes from sequence-based data obtained by STRait Razor 3.0. To demonstrate the utility of the MPS panel and data analysis workflow scheme, we also analyzed 56 microhaps of 286 samples from four populations (African-American, Caucasian, Hispanic, and Korean). The average effective number of alleles (Ae) for the four groups was 3.45, ranging from 1.74 to 6.98. Forensic statistical parameters showed that this microhap panel is more powerful than conventional autosomal STRs for human identification. Meanwhile, the 56-plex panel mostly comprised microhaps with high Ae; however, the four populations were grossly distinguishable from each other by cluster analysis. Consequently, the developed in-house MPS panel for 56 microhaps and the adopted workflow using open-source tools can increase the utility of microhap MPS in forensic research and practice.

Abstract 2945: Comprehensive analysis of tissue and plasma-related genetic alterations in Hungarian colorectal cancer patients

Abstract Background: Analysis of circulating cell-free DNA (cfDNA) of colorectal cancer (CRC) patients provides an ideal approach to explore genetic alterations in a minimally invasive way. Aims: We aimed to perform a comprehensive analysis of tissue-originated genomic DNA and plasma-derived cfDNA from CRC patients by whole-exome and targeted panel sequencing. Materials & methods: DNA was isolated from tissue and plasma samples of 55 [7 healthy (N), 16 adenomas (AD), and 32 CRC] patients using the High Pure PCR Template Kit (Roche) and the Quick cfDNA Serum & Plasma Kit (Zymo). cfDNA quality was assessed with the High Sensitivity DNA kit on BioAnalyzer 2100 and quantified by the Qubit dsDNA HS assay. Exome libraries from tissue samples were constructed by the Nextera DNA Exome Kit (Illumina). For the cfDNA samples, we used the QIAseq cfDNA All-in-One kit combined with the QIASeq Human Exome Kit, and in the case of 12 patients, CRC-specific oncogenes were further analyzed with a solution developed by QIAGEN for reliable calling of low frequency variants in cfDNA samples. Exome libraries were quantified with the KAPA Library Quantification Kit and the QIAseq™ Library Quant Assay Kit (Qiagen) and were sequenced using the NextSeq 500/550 High Output v2 kit on a NextSeq 500 Instrument (Illumina). Raw data analysis and demultiplexing were completed on the BaseSpace Sequence Hub. Variants of the tissue and cfDNA samples were determined by the Mutect2 and Haplotype caller algorithms of GATK 4.1.4.1, respectively. Clinical significance was evaluated according to the OncoKB database. Results: The most frequently mutated genes were APC, KRAS, DYNC1H1, KCNO5, and MARCH6 in the colorectal adenoma tissue samples, while those in CRC samples were APC, TP53, TTN, KRAS, and DYNCC2H1. CfDNA quantity was significantly higher in the CRC group compared to the AD (p<0.02) and N (p<0.005) patients. Based on the plasma exome results, 4.4-59.4% of the tumor somatic variants could be found in 12.5% of the patients. These patients were above 60 years of age and had Dukes D stage CRC. Panel sequencing of the circulating DNA samples detected tumor somatic variants in 8 out of the 12 enrolled patients. This method could identify 60% of all tumor somatic variants falling on its targeted regions, while whole-exome sequencing recovered only 20% of tumor somatic variants in the respective regions in cfDNA of the same patients. Conclusion: We have performed a comprehensive genetic analysis on CRC tissue and cfDNA samples in the Hungarian population. Exome sequencing offers a broad overview of the coding regions, however, targeted panel sequencing with a higher coverage depth can detect tumor somatic variants more reliably in cfDNA, therefore, it can hold a relevant clinical potential. Citation Format: Alexandra Kalmar, Gitta Szabo, Orsolya Galamb, Barbara Kinga Bartak, Zsofia Brigitta Nagy, Sara Zsigrai, Krisztina Andrea Szigeti, William Kothalawala, Peter Igaz, Istvan Takacs, Bela Molnar. Comprehensive analysis of tissue and plasma-related genetic alterations in Hungarian colorectal cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2945.

Abstract 165: Enhanced processing of genomic sequencing data for pediatric cancers: GPUs and machine learning techniques for variant detection

Abstract Introduction: The evolution of precision oncology will demand the integration of multi-omics, clinical data, demographics and outcomes, which is a complex and computationally intensive endeavor. Accurate and robust molecular characterization of pediatric leukemias require a comprehensive approach involving longitudinal sample collection across a large cohort. Genomic alterations, both germline and somatic, need to be efficiently captured and annotated to identify clinically meaningful trends. Algorithms developed for conducting genomic analysis typically leverage a central processing unit (CPU) environment for data processing and traditional statistical approach, such as Bayes' theorem, for variant detection. More recently, algorithms have been developed that can leverage a graphics processing unit (GPU) architecture and machine learning (ML) technology. In this study we benchmarked these techniques and applied them sequencing data generated from pediatric AML subjects enrolled on the Children's Oncology Group AAML1031 trial (951 samples). Methods: Whole genome sequencing was conducted using the Illumina platform (2X150) and sequenced between 30x-60x coverage. Genome alignment files were generated and subjected to germline and somatic variant calling (NVIDIA's Selene DGXA100 cluster; Parabricks Pipelines). Germline variants were called using Haplotype caller and DeepVariant (Parabricks v3.0). HG002 genome in a bottle sample was used as a positive control to determine precision, recall and F1 score. Somatic variant calling was completed using somatic sniper, VarScan2, Muse, MuTect2, and Strelka (Parabricks v3.5) and SEQCII data was used to assess somatic variant calling performance. Results: Control sample HG002 was used for benchmarking the algorithms tested. At all depths analyzed, 50X, 30X, 15X, and 5X, GoogleDeep Variant had the highest F1 score, precision, and recall compared to Haplotyper for SNPs and InDels. For somatic SNV detection using SEQCII dataset, Strelka2 had the highest F1 score, precision, and recall (0.9435, 0.9445, 0.9425), followed by MuTect2, MUSE, VarScan2, and SomaticSniper. For somatic InDel detection, MuTect2 and Strelka2 performed similar with MuTect2 higher precision, but Strelka2 had higher recall. These pipelines were utilized to analyze 951 pediatric AML samples, and demonstrated superior processing time and identification of variants, both germline and somatic. Discussion: ML algorithms integrated into a GPU computing environment demonstrated greater precision and accuracy for germline variant detection compared to traditional statistical approaches for variant detection. Ongoing efforts are focused at integration of multiple data sources correlated with outcomes and analyzed via ML tools in order to improve future risk stratification and therapeutic selection for superior outcomes. Citation Format: Erin L. Crowgey, Pankaj Vats, Karl Franke, Gary Burnett, Ankit Sethia, Timothy Harkins, Todd E. Druley. Enhanced processing of genomic sequencing data for pediatric cancers: GPUs and machine learning techniques for variant detection [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 165.

Molecular Epidemiology of Acute Myeloid Leukemia (AML): Novel Association of Clinical Epidemiologic Exposures with Unique Mutation Groups on Exome Sequencing of Leukemia DNA in the Mayo Clinic AML Epidemiology Cohort

Background: Clinical epidemiologic risk factors associated with leukemia development are prevalent, and can be identified in over 90% of AML patients. We previously demonstrated an association of clinical exposures with cytogenetic risk group and outcome after therapy, but the molecular impact of such exposures is unclear, and specifically it is not known whether specific epidemiologic exposures are associated with unique genetic lesions. We therefore performed Exome Sequencing in 98 patients from our clinical epidemiology cohort to determine whether there was an association of clinically relevant AML risk factors and exposures with unique deleterious non-synonymous (nonsyn.) mutations.Methods: The Mayo Clinic AML Epidemiology cohort comprises a highly annotated series of n=295 unselected and consecutive patients diagnosed and treated at Mayo Clinic Florida & Arizona [Cancer Epidemiol 39:1084, 2015]. We obtained AML DNA from remnant diagnostic cytogenetic pellets identified in 98 patients in the cohort, and performed Exome Sequencing [ExSeq] using standard techniques. The ~100 million paired-end 100bp reads from each sample were mapped to human genome build hg19 using BWA-MEM 1. The single nucleotide variants (SNVs) and small INDELs in the exome capture regions were identified using Mayo Clinic in-house developed analytic pipeline GenomeGPS 4.0.1 following Broad GATK 2 variant discovery best practices of alignment, realignment and recalibration, haplotype caller, and multi-sample joint genotyping. The variants passed the GATK's Variant Quality Score Recalibration (VQSR) quality control were kept. Since no patient matched normal were sequenced, SNVs and INDELs were filtered using both germline whole exome and whole genome sequencing of ~1200 Mayo Clinic Biobank samples and the public germline variant databases of 1000 genome project, 6500 individuals sequenced in the exome sequencing project (ESP), and HapMap phase 3. The remaining variants were annotated using ANNOVAR 3, and potential functional variants of non-synonymous, truncating, frame-shift, and splice-sites were used in the statistical association analyses. We evaluated the association of unique mutations occurring in 10 or more patients and in TP53, and also gene categories according to algorithm of Lindsley et al (Blood 125:1367, 2015) with relevant clinical and epidemiology exposures identified in our prior analysis. Risk factors and exposures were compared between patients with and without a mutation in different genes of interest using Fisher's exact test.Results: The read quality was excellent, and results from 97 evaluable samples were included in analysis. Results are noted in Table. We found a strong and significant association of prior history of immunosuppressive therapy (IST) or solid organ transplantation (SOT) with RUNX1 mutation (p=0.0001), and of regular Statin use with ASXL1 mutation (p=0.008). Statin use was also associated with mutations in genes involved in DNA methylation (p=0.036) and in Chromatin Modifier genes (p=0.049). We identified a novel inverse association of obesity - the most prevalent AML risk factor - with spliceosome mutation (p=0.010), and of IST/SOT with mutations in Myeloid Transcription Factors (TF, p=0.015) and Activated Signaling pathways [RAS, FLT3, KIT, CBL, NF1] (p=0.026). We confirm the association of secondary AML [sAML] with spliceosome mutation (p=0.018), observed as expected most strongly for Antecedent Hematologic Disorder (p=0.008, not shown). Other borderline associations (p≤0.10) are noted in Table. We did not identify any association of family history [FHx] of hematologic malignancy or of smoking [TOB] with unique mutations, nor did we identify clinical exposures associated with TP53 mutation.Conclusion: For the first time (apart from sAML) we have identified a unique leukemia genotype associated with some common clinical and epidemiologic exposures. Some mutations groups (Cohesin, TP53) were not associated with any clinical exposures, suggesting alternate mechanisms of leukemogenesis. ExSeq was for the first time successfully performed on archived diagnostic cytogenetic cell pellets, we believe a novel source of leukemia DNA. These results will guide a planned prospective study to determine frequency and mechanism of leukemogenesis after exposures, and development of strategies for prediction of leukemia risk and ultimately prevention. [Display omitted] DisclosuresCerhan:Janssen: Other: Scientific Advisory Board (REMICADELYM4001); Janssen: Other: Multiple Myeloma Registry Steering .

Abstract PD10-02: The influence of pathogenic variants in breast cancer predisposition genes on secondary breast cancer events in a prospectively collected cohort of breast cancer patients

Abstract Background: About 10-20% of breast cancers (BC) are hereditary. Patients (pts) with BC and pathogenic variants (PVs) in the BRCA1 or BRCA2 gene are at increased risk of contralateral breast cancer (CBC). However, risks of CBC and recurrent BC associated with PVs in other established BC predisposition genes (e.g. PALB2, CHEK2, ATM) are less well established. We performed research genetic testing of over 4000 pts with BC treated at Mayo Clinic. Because most pts were unaware of their mutation status prior to selection of therapy, this cohort provides a unique opportunity to evaluate the natural history of BC recurrence after treatment of the index BC in pts with a deleterious mutation in predisposition genes. Methods: Women diagnosed and treated for BC at Mayo Clinic between 2001-2016 were enrolled in an IRB-approved research cohort. Pts were not selected for age of diagnosis or family history. Coding regions and consensus splice sites of established BC predisposition genes were subjected to amplicon-based deep sequencing using the QIAseq method in germline DNA extracted from peripheral blood or buccal samples. PVs in the genes were identified using GATK haplotype caller and Vardict. Patient, tumor, oncologic management and follow up data were reviewed. Herein we report on the differences in the incidence of local recurrence and CBC, among carriers and non- carriers and by predisposition gene using the cumulative incidence estimator and Fine and Gray regression to account for competing risks. Results: Of 4272 women with BC (median age 57, range 21-94), a PV in a BC predisposition gene was identified in 287 (6.7%). The most common PVs were seen in CHEK2 (67, 1.6%), BRCA1 (62, 1.5%), BRCA2 (60, 1.4%), ATM (47, 1.1%); 8 (0.2%) patients had more than one PV. Pts with any PVs presented at younger age (median 51 vs 58 years, p<0.001), were more likely to have stage IV disease (4.5% vs 2.3%, p=0.02), estrogen receptor (ER) negative BC (19.5% vs 13.5%, p=0.005), and bilateral cancer (8.8% vs 4.3%, p<0.001). ER- BC proportion varied by gene - 38.9% in pts with PVs associated with ER- BC (BRCA1, BARD1, BRIP1, RAD51C, RAD51D) and 11.9% in those with PVs associated with ER+ BC (BRCA2, ATM, CDH1, CHEK2), p<0.001. NCCN criteria for genetic testing were assessed for 3763, with 69.9% of those with PVs qualifying for testing compared to 46.3% of those without PVs (p<0.001). Among 3973 pts (249 with PVs, 3724 without) with unilateral stage 0-III cancer, 168 pts developed local recurrence, 131 pts developed CBC and 248 pts developed distant disease with median 5 years of follow-up. Patients with any PV had higher rates of local recurrence after breast conserving surgery (HR 2.3, 95% CI: 1.2-4.2, p=0.01) but overall did not have significantly higher risk of CBC if contralateral prophylactic mastectomy was not performed (HR 1.7, 95% CI: 0.9-3.2, p=0.11). Specifically, pts with PVs in genes other than BRCA1 or BRCA2 did not have a higher risk of CBC compared to pts with no PV (HR 1.0, 95% CI: 0.4-2.7, p=0.96); the comparison of CBC risk in BRCA1 or BRCA2 versus those with no PV was HR 2.9 (95% CI: 1.3-6.5, p=0.01), while BRCA1 or BRCA2 compared to those with other PVs was HR 2.7 (95% CI: 0.8-9.3, p=0.13). In pts with CHEK2 PVs, the probability of CBC at 10-years follow-up was 5.4%, which was similar to pts without a mutation (5.3%), p=0.47. Conclusion: While PVs in known BC predisposition genes increase risk of BC development, risk of CBC in carriers of PVs in non-BRCA predisposition genes with BC are not as high as seen with BRCA1 or 2 PV carriers. This information regarding future BC will be helpful in personalizing decisions on the extent of surgery in carriers of PVs in BC predisposition genes. Citation Format: Judy C. Boughey, Tanya L. Hoskin, Siddhartha Yadav, Fergus J. Couch. The influence of pathogenic variants in breast cancer predisposition genes on secondary breast cancer events in a prospectively collected cohort of breast cancer patients [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD10-02.

Multithreaded variant calling in elPrep 5.

We present elPrep 5, which updates the elPrep framework for processing sequencing alignment/map files with variant calling. elPrep 5 can now execute the full pipeline described by the GATK Best Practices for variant calling, which consists of PCR and optical duplicate marking, sorting by coordinate order, base quality score recalibration, and variant calling using the haplotype caller algorithm. elPrep 5 produces identical BAM and VCF output as GATK4 while significantly reducing the runtime by parallelizing and merging the execution of the pipeline steps. Our benchmarks show that elPrep 5 speeds up the runtime of the variant calling pipeline by a factor 8-16x on both whole-exome and whole-genome data while using the same hardware resources as GATK4. This makes elPrep 5 a suitable drop-in replacement for GATK4 when faster execution times are needed.

Abstract 3554: Genomic landscape of acquired uniparental disomy in NCI PDMR patient derived xenograft models

Abstract Background: Acquired Uniparental Disomy (aUPD) is relatively common in cancer. Occurrence of aUPD is more frequent in some tumor histologies (e.g., serous ovarian, colorectal) and may be relevant for choice of therapy. The Patient-Derived Models Repository (PDMR; https://pdmr.cancer.gov) developed by The National Cancer Institute (NCI) includes patient-derived xenograft (PDX) models from multiple tumor histologies with different passages and lineages. The associated clinical annotation and genomic data make it possible to assess the prevalence of aUPD in the PDMR cohort and the stability of aUPD in different passages and lineages within a PDX model. Methods: High tumor purity in the PDX specimens (after removal of mouse reads representing the stroma) enabled highly accurate assessment of loss of heterozygosity (LOH). Variants called by GATK Haplotype caller from whole exome sequencing (WES) data were used to identify segments of homozygosity using BCFtools/RoH (runs of homozygosity). The RoH segments were then intersected with the bed file for chromosome arms to get %LOH at the arm level. If %LOH on a chromosome arm was >90%, we considered the sample to have aUPD at the arm level. WES was also used to look for associations between DNA damage repair (DDR) pathway alterations and aUPD. Results: We made the following observations: a) aUPD was observed most frequently in chr18q (75/427, 17.6%) and chr3p (69/427, 16%) of PDX models; b) aUPD was observed more frequently in certain tumor histologies, e.g., clear cell renal cell carcinoma (6/8), small cell lung cancer (3/4) and non-small cell lung cancer (25/38); c) extensive aUPD was observed in 4 PDMR models (>50% of evaluated chromosome arms in these models have aUPD); d) aUPD was not observed in some tumor histologies, i.e., synovial sarcoma, uterine endometrioid carcinoma; e) in the vast majority of PDMR models (>90%), aUPD is maintained faithfully across lineages and through multiple passaging; f) subclonal aUPD events were observed in some models across different lineages; g) significant enrichment of double strand DNA break repair (DSBR) pathway alterations was observed in PDMR models without aUPD (p=0.0007, Fisher's exact test) suggesting defects in DSBR are not associated with aUPD; and h) aUPD was rarely observed in MSI-high models (1/30) suggesting mutual exclusivity of mismatch repair (MMR) pathway defects and aUPD. Conclusion: We observed a relatively high frequency of UPD in the PDMR models (at least 1 arm of a chromosome). UPD was more frequently observed in specific chromosomal arms. The frequency of aUPD was higher in some tumor histologies and absent in others. aUPD was stably maintained across passages and lineages, although some heterogeneity was observed. Our data suggest aUPD is not associated with defects in DSBR and MMR pathways. Preclinical drug studies using NCI PDMR models may suggest appropriate therapeutic options for cancers with aUPD. Citation Format: Rajesh Patidar, Li Chen, Chris A. Karlovich, Biswajit Das, Yvonne A. Evrard, Tomas Vilimas, Justine N. McCutcheon, Amanda L. Peach, Nikitha V. Nair, Thomas D. Forbes, Brandie A. Fullmer, Anna J. Lee Fong, Luis E. Romero, Alyssa K. Chapman, Kelsey A. Conley, Robin D. Harrington, Shahanawaz S. Jiwani, Peng Wang, Michelle M. Gottholm Ahalt, Erin N. Cantu, Gloryvee Rivera, Lindsay M. Dutko, Kelly M. Benauer, Vishnuprabha R. Kannan, Suzanne D. Borgel, John P. Carter, Jesse M. Stottlemyer, Tiffanie L. Miner, Devynn R. Breen, Emily T. Delaney, Chelsea A. McGlynn, Candace N. Mallow, Marianne Radzyminski, Shannon N. Uzelac, Sergio Y. Alcoser, Tara L. Grinnage-Pulley, Michelle A. Eugeni, Dianne L. Newton, Melinda G. Hollingshead, Paul M. Williams, James H. Doroshow. Genomic landscape of acquired uniparental disomy in NCI PDMR patient derived xenograft models [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 3554.

GPU accelerated sequence alignment with traceback for GATK HaplotypeCaller

BackgroundPairwise sequence alignment is widely used in many biological tools and applications. Existing GPU accelerated implementations mainly focus on calculating optimal alignment score and omit identifying the optimal alignment itself. In GATK HaplotypeCaller (HC), the semi-global pairwise sequence alignment with traceback has so far been difficult to accelerate effectively on GPUs.ResultsWe first analyze the characteristics of the semi-global alignment with traceback in GATK HC and then propose a new algorithm that allows for retrieving the optimal alignment efficiently on GPUs. For the first stage, we choose intra-task parallelization model to calculate the position of the optimal alignment score and the backtracking matrix. Moreover, in the first stage, our GPU implementation also records the length of consecutive matches/mismatches in addition to lengths of consecutive insertions and deletions as in the CPU-based implementation. This helps efficiently retrieve the backtracking matrix to obtain the optimal alignment in the second stage.ConclusionsExperimental results show that our alignment kernel with traceback is up to 80x and 14.14x faster than its CPU counterpart with synthetic datasets and real datasets, respectively. When integrated into GATK HC (alongside a GPU accelerated pair-HMMs forward kernel), the overall acceleration is 2.3x faster than the baseline GATK HC implementation, and 1.34x faster than the GATK HC implementation with the integrated GPU-based pair-HMMs forward algorithm. Although the methods proposed in this paper is to improve the performance of GATK HC, they can also be used in other pairwise alignments and applications.

SA10 - USING WHOLE GENOME SEQUENCING TO IDENTIFY DE NOVO VARIATION IN BIPOLAR DISORDER

Background Although most gene-discovery efforts in Bipolar Disorder (BD) have focused on inherited variation, the association of BD with increased paternal age and prior association of de novo copy-number variants with BD, suggests that de novo mutations may also play a role. Methods We have performed deep whole genome sequencing of 97 trios selected specifically to have no family history of bipolar disorder. Diagnoses were based on the Diagnostic Instrument for Genetics Studies with two independent best-estimate evaluations. Sequencing was performed on Illumina Hi-Seq with an average depth of 37X. Variant calling was performed according using the GATK using haplotype caller and best-practice recommendations. De novo calls were individually visualized with the Integrative Genomics Viewer and a subset, including all exonic variation, underwent confirmatory Sanger sequencing. Results We identified 6,882 de novo variants throughout the genome, representing an average of 73.3 de novo Single Nucleotide Variants (SNVs) and indels per family (range 43 -106). As expected, the number of de novo variants correlated strongly with paternal age (P Discussion Our whole genome study of de novo variation provides initial evidence for an association of disruptive mutations in highly conserved genes with BD. The consistency of our results with the only other study of de novo variation in BD highlights the need for further study of de novo variation in BD.

Targeted next-generation sequencing (NGS) of germline DNA to identify genetic predisposition to gastric cancer (GC) in patients with CDH1-mutation negative early-onset (EO) or familial GC (FGC).

20 Background: While GC arises as part of several hereditary cancer syndromes, a genetic etiology is not identified for most FGC kindreds and EOGC patients. We hypothesize that NGS of germline DNA may reveal previously unsuspected genomic alterations that predispose to EOGC/FGC. Methods: We identified 42 Pts from a prospective GC registry with CDH1-mutation negative EOGC (≤50 years old) and/or with FGC. Germline DNA was analyzed by MSK-IMPACT, a 468-gene targeted NGS panel that includes > 70 genes associated with known cancer predisposition syndromes. GATK HaplotypeCaller (HTC) was used to call SNVs and Indels simultaneously from the sequence read data and joint calling of all samples was used to generate VCF files. We annotated the VCF with public annotators and public allele frequencies from large sequencing studies. Manual variant curation was performed and were classified by ACMG criteria (Genet Med 2015;17:405), with only likely pathogenic and pathogenic variants included. Results: 41 Pts (98%) had early onset and 11 Pts (26%) had FGC (9 had ≥1 1st or 2nd degree relatives with GC). Median age was 40 (range, 18 to 55) and 23 (55%) were men. Ethnicities were 21 (50%) White, 13 (31%) Black, 6 (14%) Asian and 2 (5%) unknown. 40 (95%) had poorly differentiated and/or diffuse histology. Pathogenic mutations were diagnosed in 4 Pts (10%). 1 Pt had a BRCA2 mutation; his father had prostate CA and a paternal aunt had breast CA but he did not meet current NCCN criteria for BRCA testing. Three additional pathogenic mutations were FGFR1, SOX17 and KMT2D; none has previously been described in GC susceptibility but are associated with GC carcinogenesis. All 3 Pts had EOGC. None had FGC but all 3 had ³1 1st-degree relative with CA ( FGFR1: breast, uterine, appendix; SOX17: breast; KMT2D: osteosarcoma). Further functional work of these variants and segregation analysis is on-going. Conclusions: MSK-IMPACT identified germline mutations in 10% of Pts; most are previously not known GC predisposition genes. NGS can be an important clinical tool for screening multiplex families and also a research tool for discovering genes that predispose to GC.

Efficient Acceleration of the Pair-HMMs Forward Algorithm for GATK HaplotypeCaller on Graphics Processing Units.

GATK HaplotypeCaller (HC) is a popular variant caller, which is widely used to identify variants in complex genomes. However, due to its high variants detection accuracy, it suffers from long execution time. In GATK HC, the pair-HMMs forward algorithm accounts for a large percentage of the total execution time. This article proposes to accelerate the pair-HMMs forward algorithm on graphics processing units (GPUs) to improve the performance of GATK HC. This article presents several GPU-based implementations of the pair-HMMs forward algorithm. It also analyzes the performance bottlenecks of the implementations on an NVIDIA Tesla K40 card with various data sets. Based on these results and the characteristics of GATK HC, we are able to identify the GPU-based implementations with the highest performance for the various analyzed data sets. Experimental results show that the GPU-based implementations of the pair-HMMs forward algorithm achieve a speedup of up to 5.47× over existing GPU-based implementations.

Germline and somatic SNVS calling in NGS panel tumor samples: Approaches to optimize tumor only genomic analysis for cancer precision medicine.

e13011 Background: Clinical NGS is often limited by tumor only profiling. Discrimination between somatic and likely germline mutations when calling from tumor patient samples is a critical step for clinical genotyping. Many algorithms have been developed for somatic single nucleotide variant (SNV) detection in matched tumor-normal whole genome and whole exome sequencing. Here, we demonstrate approaches of how a cost-effective large gene panel sequencing can be used to call somatic and germline SNVs for tumor only samples. Methods: Tumor, adjacent normal, and matched normal samples are collected from five patients. The somatic mutations were called with GATK Mutect2 in tumor only and adjacent normal. The germline mutations were called individually for all 15 samples with GATK Haplotype caller. To remove germline mutations from tumor only somatic calls, the filters ExAc pop freq, 1000G pop freq, COSMIC were applied on the tumor only somatic calls. PPV (Positive predictive value) for each filter was calculated by dividing the number of somatic mutations in the post-filtering mutation data by the total number of unfiltered mutations. TPR (True Positive Rate, representing sensitivity) was calculated by dividing the number of true somatic mutations in the tumor-only post-filter. Results: Compared with germline mutations called from matched normal, 70% germline mutations were called in RAWE somatic calls. A PPV of 0.71 and a TPR of 0.95 were optimally provided when the filter ExAc pop freq > 0.01 and COSMIC ( > 5 occurrence) applied. For germline mutations called in tumor samples, when compared with those in blood samples and in adjacent normal samples, PPV is 0.99 and TPR is 0.97. For somatic mutations called with tumor-adjacent normal pair mode in Mutect2, PPV is 0.5 and TPR is 0.99. Conclusions: Optimization of tools and parameters in NGS large panels could detect somatic and germline variants with high specificity, sensitivity and accuracy, without matched or adjacent normal. For the germline variants, when adjacent normal is available, it could replace matched normal with high accuracy.

M26 - Whole Genome Sequencing To Identify De Novo Mutations In Bipolar Disorder

Bipolar Disorder (BD) is a common, complex disorder that ranks among the leading causes of disability worldwide. Based on its prominent heritability, most gene discovery efforts have focused on inherited variation. However, the association of BD with increased paternal age and prior association of de novo copy-number variants, suggests that de novo mutations may also play a role. As de novo mutations have not undergone negative selection, they may represent a potentially highly penetrant class of variants particularly useful for highlighting illness-related biology. We have performed deep whole genome sequencing of 97 simplex trios of Ashkenazi descent that were selected to have no family history of Bipolar Disorder or Schizophrenia. Probands were diagnosed with Bipolar Disorder type I based on the Diagnostic Instrument for Genetics Studies (DIGS) with two independent best-estimate evaluations. Sequencing was performed on Illumina Hi-Seq with an average depth of 37X. Variant calling was performed using the GATK using haplotype caller and best-practice recommendations. De novo calls were individually visualized with the Integrative Genomics Viewer and a subset, including all exonic variations, underwent confirmatory Sanger sequencing. We identified 6,186 de novo variants throughout the genome, representing an average of 63.7 de novo single nucleotide variants (SNVs) and indels per family (range 40 -97). The number of de novo variants correlated strongly with paternal age (P<4.8 x10-14), with an average increase of 1.8 mutations per year of age. We identified 105 de novo coding mutations (average 1.06/trio) at rates consistent with Poisson expectations. Among these, there were 12 de novo disruptive variants, including 6 disruptive variants in highly conserved genes, which represents a nominal enrichment (P = 0.038) compared to null expectations. Incorporating a recently published exome-based de novo study of BD provided further evidence for enrichment (P=0.00594) of disruptive mutations in highly conserved genes. Ongoing analysis of de novo variants in the non-coding genome will also be presented. Our ongoing whole genome study of de novo variation shows that, consistent with other psychiatric disorders, there is evidence for an association of disruptive mutations in highly conserved genes with BD. Although no recurrent mutations were found in the same gene, the consistency of our results with the only other study of de novo variation in BD highlights the need for further study of de novo variation in BD

Using Whole Exome Sequencing to Identify Genetic Variation and Polymorphisms Associated with Graft Versus Host Disease in Allogeneic Stem Cell Transplant Recipients

Objective: Allogeneic stem cell transplant (ASCT) is a potentially curative treatment for many hematologic malignancies. HLA match is the principal factor in donor selection but the total antigenic landscape as seen by the donor immune cells extends beyond the HLA and minor HLA antigens. Thus, despite "perfect" HLA match graft versus host disease (GVHD) caused by alloreactive donor immune cells reacting to disparate antigens in the recipient, remains a significant cause of morbidity and mortality. Whole exome sequencing (WES) which assays all the protein coding nucleotide sequences provides a powerful means to assess the entire antigenic makeup of the individual. We examined the WES of sibling donor-recipient pairs undergoing BMT to assess the influence of the extent of genetic disparity and specific gene polymorphisms on GVHD.Methods: Sixteen matched sibling donor-recipient pairs receiving allogeneic stem cell transplants between June 2008 and April 2013 had samples available for retrospective WES. Transplant data including conditioning regimen, age of donor and recipient at the time of transplant, development and type of GVHD and cause of death were recorded. The sequencing was performed using Illumina HiSeq2000 and the single nucleotide variants (SNVs) and small insertions and deletions (INDELs) were identified using the haplotype caller within the Genome Analysis Toolkit GATK v2.7. Functional variants including non-synonymous, frame-shift, truncating, and splice site SNVs and INDELs identified were compared between donor and recipient and among those with GVHD and those without GVHD (non-GVHD). One-sided Fisher's exact test was used to determine the statistical significance of the SNPs in GVHD and non-GVHD. Candidate genes with a p value of <0.05 were further evaluated for their biochemical pathway and possible role in inflammatory or immune responses.Results: Nine of 16 patients had a diagnosis of AML or high grade myelodysplastic syndrome, 3 had a diagnosis of non- Hodgkin lymphoma, 1 had Hodgkin lymphoma, 1 had multiple myeloma, and 2 had CML. Six patients were conditioned with a myeloablative regimen, others with a reduced intensity conditioning regimen. Eight of 16 patients were female and 7 of 16 donors were female. Nine of 16 patients developed stage II-IV acute and/or chronic GVHD with 7 of them developing GVHD of the skin. A total of 12,623 genes with 43,670 functional variants were identified using WES. 134 genes had a p value < 0.05 and enrichment analysis was performed on these genes. 12 out of 134 genes were associated with gene ontology (GO) term "immune system development" with an unadjusted enrichment p-value 0.0015. Three (IL31RA, OSMR, SIRPA) out of 134 genes were associated with "inflammation_Jak-STAT pathway" with an adjusted enrichment p-value 0.07. Variants of IL31RA, OSMR, and PTX3 were identified as genes with polymorphisms associated with GVHD (p= 0.003, 0.01, and 0.01, respectively) with 5 variants identified in IL31RA, 7 variants in OSMR, and 2 variants in PTX3. Further investigation of these three genes revealed their role in inflammatory or immune responses.Discussion: GVHD is a cytokine mediated immune-inflammatory process that can occur despite "perfect" HLA matching. WES identified the extensive genetic variability between donors and recipients implying large allo-reactivity potential that can lead to GVHD. Quantification of genetic variation may allow for more precise determination of GVHD risk and better selection of donors. In our cohort, WES identified polymorphisms in IL31RA, OSMR, and PTX3 as statistically significantly associated with development of GVHD. The protein products of these three genes have each been shown to be involved in inflammatory processes and are influenced by other cytokines. IL31RA and OSMR are found in many tissues including the skin and lungs. They function both independently and as a heterodimer receptor capable of binding multiple cytokines including IL-31, IL-6, and TNF-α. PTX3 is a reactive protein that increases in response to cytokine release related to systemic inflammation. Levels of this protein have been used as a prognostic marker in many disease processes with high levels correlating with poor outcomes. The small sample size of our cohort precludes any firm conclusions but provides promising hypotheses for further evaluation in larger studies. DisclosuresNo relevant conflicts of interest to declare.

Abstract 4831: Analysis of genetic variations in neuroblastoma using deep targeted sequencing

Abstract Neuroblastoma tumors (NBs) show complex patterns of genetic abnormalities, which may include amplification or oncogenic mutations of ALK kinase in sporadic and familial cases. In analyzing genomic deviations it is sometimes highly efficient to focus on deep sequencing of specific parts of the genome with high clinical relevance to NB prognosis. This can be achieved by studying specific regions using targeted or amplicon sequencing, and thereby reducing the complexity of the analysis. The main benefits of this approach are reduced costs and shorter analysis time, due to the sequencing of fewer bases and the associated reduction in the volume of data generated. Additional benefit comes from achieving higher coverage that provides means for detection tumor specific variants present at subclonal level. The HaloPlex method, the custom target enrichment kit from Agilent, uses a different approach among enrichment kits, namely the fragmentation of DNA by restriction enzymes in conjunction with biotinylated probes designed to hybridize to both ends of a targeted DNA restriction fragment. This will form circular DNA strands that will be then closed by ligation (http://www.genomics.agilent.com) and it requires only 250 ng of input material. We targeted 35kb of genomic sequence including exons of related genes, as well as 50 bp of flanking exon regions. Samples were sequenced in a single MiSeq run (Illumina) using 2×150 paired-end sequencing. We compiled a gene panel for Next Generation Sequencing (NGS) comprising of three genes, ALK, ATRX and PPM1D, where the two first are associated with different subgroups of NBs. The Haloplex panel was evaluated on a pilot cohort of 40 patients including four controls with known and defined mutations (missense and indels), which produced a total of 4 Gb of data. Data were analyzed using an “in-house” bioinformatics pipeline based on mapping reads against hg19 using bwa, with realignment and recalibration using GATK haplotype caller, and variant annotation using ANNOVAR. For each position, the frequency for each base was calculated. The analysis revealed that we were able to generate sufficient sequencing data with very low background variability for mutation as well as deletion analysis using our diagnostic panels. We identified all expected mutations in our control samples. Sanger sequencing were used to validate the results and for analyzing sensitivity and specificity of the detected variants. Overall, the panel fulfills diagnostic criteria and will offer a fast and reliable procedure that will be useful in clarifying molecular genetic defects underlying NBs. Citation Format: Niloufar Javanmardi, Susanne Fransson, Malin Ostensson, Rose-marie Sjoberg, Per Kogner, Tommy Martinsson. Analysis of genetic variations in neuroblastoma using deep targeted sequencing. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4831. doi:10.1158/1538-7445.AM2015-4831

Abstract PD3-2: Broad exonic DNA diversity is associated with resistance to taxane-FAC chemotherapy in triple negative breast cancer

Abstract Purpose: Previous efforts to develop transcriptional markers of chemotherapy sensitivity in TNBC had limited success due to the heterogeneity of this disease. The purpose of this study was to identify genomic differences between extremely chemotherapy sensitive and highly chemotherapy resistant TNBC through whole exome sequencing and to assess measures of genomic heterogeneity as predictive markers. Methods: Twenty nine cases were selected from a prospectively collected cohort of fine needle aspiration specimens obtained before preoperative chemotherapy (MDACC) to represent two extreme response cohorts including pathologic complete response (pCR, N=17) or extensive residual cancer (eRD, N=12). DNA was extracted from specimens stored in RNAlater, exomes were captured using NimbleGen SeqCap EZ Exome Library preparation and paired-end sequencing of 75 base pair fragments was performed on Illumina HiSeq 2000. Alignment and variant calling were performed with BWA and GATK haplotype caller. Variants were filtered against the 1000 Genomes and TCGA normal breast samples to identify candidate somatic variants. Fisher-exact test was used to identify variant genes associated with sensitivity to chemotherapy. We calculated overall mutational load and normal-adjusted clonal entropy of driver mutations as measures of genome heterogeneity. The chi-squared test was used to compare differences in mutational spectra and genome heterogeneity between the two response groups. Results: The mean coverage was over 150X and 93% of target regions had &amp;gt; 20X coverage. The number of non-silent COSMIC mutations was similar in tumors from the two response groups (pCR: 63, range 49-82; eRD: 59, range 43-78) as well as the number of novel non-silent mutations (eRD: 223, range 113-388; pCR: 192, range 125-293). Gene level aggregation of variants identified 4 genes (MUC21, SLCO5A1, LRBA, STNE1) with response-associated mutational patterns. However, mutations were non recurrent and p-values were modest, ranging from 0.04 to 0.005. We observed greater overall mutational load and subclonal heterogeneity (clonal entropy of cancer related mutations) associated with eRD compared to pCR. Both measures suggest that higher genomic DNA diversity is associated with chemotherapy resistance. However, some genes (BRCA1 and MKI67) had higher mutational load (sum of minor allele frequencies per gene) associated with pCR compared to eRD. In general, a higher proportion of C&amp;gt;T transition (P=0.011) and lower A&amp;gt;G transition (P=0.028) was associated with eRD. The same mutational spectrum shift was previously described in the comparison of trunk and branch driver events suggesting that eRD tumors may undergo heterogeneous branched evolution. Conclusion: We observed greater genomic diversity and distinctive mutational spectra in original pre-treatment samples of TNBC that were associated with extensive residual disease compared to pCR. Our analysis suggests that broad measures of genomic diversity may serve as markers of resistance to chemotherapy. Citation Format: Tingting Jiang, Weiwei Shi, William F Symmans, Charles Li, James Platt, Rosanna Lau, Vikram B Wali, Richard Lifton, Lajos Pusztai, Christos Hatzis. Broad exonic DNA diversity is associated with resistance to taxane-FAC chemotherapy in triple negative breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr PD3-2.

Abstract PD3-4: Reliability of whole exome sequencing for assessing intratumor heterogeneity from breast tumor biopsies

Abstract BACKGROUND: False positive findings introduced by analytical noise in sequencing and bioinformatics pipelines constitute a challenge for accurate massively parallel sequencing (MPS). Reports of intratumor genomic heterogeneity based on MPS rarely estimate the impact of false positive mutation calls. The purpose of this study was to measure apparent genomic heterogeneity in different regions of the same tumor and to assess the technical noise in variant calling in replicate sequencing of the same DNA. METHODS: Three anatomically distinct biopsies were obtained from 3 different regions of 11 breast cancers (33 samples) including 6 low/intermediate grade, estrogen receptor (ER)-positive and 5 high-grade, triple-negative (TNBC) cancers. DNA from 8 different biopsies was split and independently processed on different days to obtain technical replicates. The NimbleGen SeqCap EZ Exome Library preparation method was used for exome capture and paired-end sequencing of 75 base pair fragments was performed on Illumina HiSeq 2000. Read alignment and variant calling were performed with BWA and GATK haplotype caller. Concordance in variant calls and in minor allele frequencies (MAF) was assessed in the 3 biopsies of the same tumor and 8 technical replicates. We adjusted for uneven sequence coverage and analyzed known germline variants from dbSNP, known cancer related variants from COSMIC separately from novel variants (i.e. not previously reported in dbSNP or COSMIC). We estimated intratumor genomic heterogeneity of genes after removing alterations identified in areas where mapping is difficult and variant calls that had low analytical reliability in the technical replicates. RESULTS: The mean coverage was over 150X and &amp;gt; 90% of target regions had ≥ 20X coverage. We validated the specificity (98.2%) and sensitivity (86.7%) of the variant calling pipeline on the GIAB reference data. The concordance for germ line SNPs and variants in COSMIC in technical replicates was 94.9% and 92.7%, respectively. Novel variants had very low concordance, 55.9%, in technical replicates. The concordance between MAF estimates from the technical replicates was high (0.974, 0.957 and 0.969 for single nucleotide variations, insertions and deletions, respectively). The concordance for germline SNPs and COSMIC variants in pairwise comparisons of biopsies from the same tumor was 93.2% and 91.1%. For known variants, lower concordance was observed in TNBC (88.3%-98.5%) compared to ER-positive tumors (93.6%-98.8%, P&amp;lt;0.05) indicating greater intratumor heterogeneity. We identified variants in a small number of genes (DNAH9, PPM1E, and MAP3K1) that were called inconsistently in most technical replicates, even after excluding low mappability regions. We assessed intratumor heterogeneity in the triplicate biopsies, after excluding the technically unreliable variants. On average, two different biopsies from the same tumor shared 14272 +/-1379 common variants and differed in 816 +/- 416 variants. CONCLUSION: We observed heterogeneity to be slightly greater in high-grade, ER-negative compared to low-grade, ER-positive breast cancers. Differences in variants observed in multiple biopsies of the same tumor are only slightly greater than those expected by technical noise alone. Citation Format: Weiwei Shi, Anees Chagpar, Tingting Jiang, Donald R Lannin, Brigid Killelea, Nina Horowitz, Raymond Lim, James Platt, Charlotte KY Ng, Vikram B Wali, Britta Weigelt, Jorge S Reis-Filho, Christos Hatzis, Lajos Pusztai. Reliability of whole exome sequencing for assessing intratumor heterogeneity from breast tumor biopsies [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr PD3-4.

The First Kazakh Whole Genomes: The First Report of NGS Data.

IntroductionThe human genome sequence will underpin human biology and medicine in the next century, providing a single, essential reference to all genetic information. Extraordinary technological advances and decreases in the cost of DNA sequencing have made the possibility of whole genome sequencing (WGS) feasible as a highly accessible test for numerous indications. The international project “Genetic architecture of Kazakh population” is well underway to determine the complete DNA. Next generation sequencing is a powerful tool for genetic analysis, which will enable us to uncover the association of loci at specific sites in the genome associated with disease. The aim of this study was to introduce first data on WGS of 6 Kazakh individuals.MethodsThis pilot study is among the first WGS performed on 6 healthy Kazakh individuals, using next generation sequencing platform HiSeq2000, Illumina by manufacturer’s protocols. All generated *.bcl files were simultaneously converted and demultiplexed using bcl2fasta application. Alignment of sequence reads performed using bwa-mem against human b19 reference genome. Sorting, removing of intermediate files, *.bam files assembling, and marking duplicates were performed using PicardTools package. GATK haplotype caller tool was used for variant calling. ClinVar, SNPedia, and Cosmic databases were processed to identify clinical genomic variants in 6 Kazakh whole genomes. Java Runtime Environment and R. Bioconductor packages were installed to perform raw data processing and run program scripts.ResultsThe sequence alignment and mapping procedures on reference genome hg19 of each 6 healthy Kazakh individual were completed. Between 87,308,581,400 and 107,526,741,301 total base pairs were sequenced with average coverage x29.85. Between 98.85% and 99.58% base pairs were totally mapped and on average 96.07% were properly paired. Het/Hom and Ti/Tv ratios for each whole genome ranged from 1.35 to 1.52 and from 2.07 to 2.08, respectively. We compared and analyzed each genome with on existing clinical databases ClinVar, SNPedia, Cosmic and found from 20 to 25, from 269 to 288, from 7 to 12 SNP records, respectively. The availability of a reference Kazakh genome sequences provides the basis for studying the nature of sequence variation, particularly single nucleotide polymorphisms.ConclusionThe first whole genome sequencing of Kazakhs were performed. In this pilot study, we identified SNPs associated with different conditions. Further studies of WGS on Kazakh population are needed to identify possible unique genetic variants in Kazakhs.