1
Yes
None
1
20
Search...
/dacc2/conference-home/2020-presentations/
Thumbnail
Sort By:

Scientific Category

Presentation Type

200 products«1 of 10»
Genomic Medicine
200 products«1 of 10»
200 products«1 of 10»
Genomic Medicine
Using Electronic Health Records and Machine Learning to Assess Opioid Addiction Risk
Philip Freda
Compare
Using Electronic Health Records and Machine Learning to Assess Opioid Addiction Risk Image

Connect with Philip @Freda-37 on Slack

Direct Message the author

BACKGROUND: The opioid epidemic continues to result in significant economic and cultural burdens worldwide. Despite clinicians prescribing less opioids, opioid-related fatal overdoses rise each year as addicted individuals seek out cheap, illicit opioids like heroin and fentanyl from non-medical sources. In response to these issues, research has focused on understanding how to treat addicted individuals by developing improvements to medical assistance treatment programs and exploring new medications in an attempt to limit relapse. Despite the benefits that this research provides, it is equally important to understand why patients become dependent in response to opioids. Opioid risk generally arises from genetics, psychiatric disorders, other substance abuse disorders, environmental stressors, or combinations of these sources. However, it is difficult to quantity the relative contributions of these factors for opioid addiction risk. Also, there are likely risk factors that are unknown to researchers. METHODS: Electronic Health Records (EHR) provide an abundance of information including past and current medications, medical diagnoses, and detailed clinical notes. Additionally, genomic biobanks provide genetic sequences for large cohorts of patients with detailed associated phenotypes. These data can be used to assess opioid addiction risk by determining factors and genomic variants that are common in individuals that form opioid dependence. Machine learning techniques provide efficient and effective methods to handle large datasets like these. SIGNIFICANCE: This work provides a framework and workflow that can be used to assess opioid addiction risk from EHR and genomic data using Natural Language Processing (NLP) and automated machine learning (Penn AI).
Understanding Impact of Influenza Infection On Bacterial Interactions Within the Nose/Throat Microbiome
Hasan Abu-Amara
Compare
Understanding Impact of Influenza Infection On Bacterial Interactions Within the Nose/Throat Microbiome Image

Connect with Hasan @Abu-Amara-189 on Slack

Direct Message the author

Background: Influenza remains a significant source of illness and mortality causing 490,600 hospitalizations and 34,200 deaths in 2018-2019. Much of the severe morbidity and mortality is due to secondary bacterial infections, particularly pneumonia. We examined the effects of influenza infection on Streptococcus pneumoniae and Staphylococcus aureus, which often cause pneumonia following influenza. Methods: We analyzed data from 224 children (<18 years old) and 313 adults participating in the Nicaraguan Household Transmission Study between 2012 and 2014. Nose/throat samples were analyzed to detect influenza, S. pneumoniae and S. aureus and relevant metadata collected. Results: The probability of acquiring S. pneumoniae among people who eventually contracted influenza was 1.56 times that of people who never caught influenza, for S. aureus it was 0.621. Acquisition of S. aureus was weakest in children, while acquisition of S. pneumoniae was strongest in adults. The probability of detecting S. pneumoniae only at follow-up given both organisms at enrollment was 2.13 times higher among those who acquired influenza, but the same ratio for S. aureus was 1.07. We could not analyze this relationship across age due to small sample size. Significance: To our knowledge, this is the first study prospectively examining S. aureus and S. pneumoniae interactions during influenza infection among different age groups. We found evidence that influenza infection may be associated with increased colonization of S. pneumoniae especially in adults. Our results support the hypothesis that influenza infection may promote growth of species associated with pneumonia in the nose/throat microbiome, possibly increasing risk of pneumonia.
Exome-by-phenome-wide association with electronic health record phenotypes
Joseph Park
Compare
Exome-by-phenome-wide association with electronic health record phenotypes Image

Connect with Joseph @Park-5 on Slack

Direct Message the author

BACKGROUND: Coupling DNA sequencing with electronic health records (EHR) can enhance our understanding of the contribution of genetic variation to human disease. Aggregating rare, loss-of-function variants in a gene into a “gene burden” to test in an unbiased manner for association with EHR phenotypes can identify novel clinical implications for the gene in human disease. However, this methodology has not yet been applied on both an exome-wide and phenome-wide scale, and the clinical impact of rare loss-of-function variants in many genes have yet to be described. METHODS: Leveraging 11,451 whole exomes linked to EHR data in the Penn Medicine Biobank (PMBB), we collapsed rare (minor allele frequency ≤ 0.1%) predicted loss-of-function (pLOF) variants per gene on an exome-wide scale and performed a phenome-wide association study (PheWAS) with diverse EHR phenotypes. We replicated significant associations from this discovery experiment using other approaches and the UK Biobank. RESULTS: We identified 106 exome-by-phenome-wide significant gene-phenotype relationships in the discovery experiment in PMBB. Of these, we replicated 29 associations, of which five represented ‘positive controls’ and 24 were novel findings. DISCUSSION: We thus show the value of aggregating rare pLOF variants into gene burdens for unbiased association with EHR phenotypes to identify novel clinical relationships for mutated human genes. We suggest that this approach applied to even larger cohorts of individuals with DNA sequence data linked to EHR-derived phenotype data will yield many new insights into the relationship between genetic variation and disease phenotypes.
Mobile Element Insertions and Associated Structural Variants in Longitudinal Breast Cancer Samples
Cody Steely
Compare
Mobile Element Insertions and Associated Structural Variants in Longitudinal Breast Cancer Samples Image

Connect with Cody @Steely-32 on Slack

Direct Message the author

BACKGROUND: Mobile elements are segments of DNA capable of mobilizing from one location to another. These elements are dispersed throughout the genome, composing roughly half of the human genome. Mobile elements have been valuable markers for disease genetics, population genetics, phylogenetics, and other evolutionary analyses. While mobile elements are largely inactive in healthy somatic tissues, increased activity has been found in cancer tissues, with significant variation among different cancer types. In addition to insertion events, mobile elements have also been found to mediate many structural variation events in the genome. METHODS: Here, to better understand the timing and impact of mobile element insertions and mobile element-mediated structural variants in cancer, we examined their activity in longitudinal samples of four metastatic breast cancer patients. With whole-genome sequencing data from multiple timepoints through tumor treatment and progression, we used mobile element detection software followed by visual confirmation of the insertions. RESULTS: From this analysis we identified 11 mobile element insertions or mobile element-mediated structural variants, and found that the majority (nine of the eleven) of these occur early in tumor progression. Most of the insertions and variants appear to impact intergenic regions; however, we identified a mobile element-mediated translocation in MAP2K4 and a mobile element-mediated deletion in YTHDF2 that likely inactivate reported tumor suppressor genes. SIGNIFICANCE: Overall, we find that most insertions are likely passenger mutations in the four patients we examined, but some variants may impact tumor progression.
Does a patient-facing family history collection tool reduce disparity in identification of individuals at risk of hereditary cancer?
Kathleen Mittendorf
Compare
Does a patient-facing family history collection tool reduce disparity in identification of individuals at risk of hereditary cancer?  Image

Connect with Kathleen @Mittendorf-58 on Slack

Direct Message the author

Background: Identifying patients with hereditary cancer syndromes requires in-depth family health history, which is rarely collected in enough detail to facilitate genetic counseling referral – a roadblock that disproportionately impacts individuals with other barriers to healthcare access. Methods: We adapted a bilingual patient-facing electronic web tool using validated risk assessment applications (B-RST and PREMM5). Adaptations included adding literacy aids (e.g., infographics, definitions) and reducing numeracy/literacy requirements. We are evaluating the success of these adaptations in two large healthcare systems with diverse populations, including participants with barriers to healthcare access such as race, ethnicity, geographic, socioeconomic status, and/or language barriers. Results: Most (93%) of the first 649 individuals to complete the tool did so without assistance; unassisted individuals spent a mean of 4.2 minutes (range: 0.4-53 min) on the assessment. In comparison to genetic counselor (GC)-collected family health history (FHH), preliminary data from the first 124 individuals receiving genetic counseling indicate that most (85% of patients) have their risk-level appropriately categorized by the patient-facing PREMM5 collection tool. Future analyses will examine accuracy of risk stratification by B-RST, evaluate statistical predictors of appropriate risk stratification and time spent on the tool, and examine whether the tool adds utility over existing FHH documented in the EMR. Data collection is ongoing; we will present additional updated analyses. Significance: We demonstrate that the tool is easy to use, rapid, and accurate. We will evaluate whether this tool in primary care can facilitate genetic referral for individuals with additional barriers to healthcare access.
MAPRE2 may be a novel Brugada syndrome gene
David Chiang
Compare
MAPRE2 may be a novel Brugada syndrome gene Image

Connect with David @Chiang-63 on Slack

Direct Message the author

Background: Brugada syndrome (BrS) is a significant cause of sudden cardiac death in young men yet only ~30% of patients has a known genetic cause. A recent GWAS found 9 novel loci associated with BrS, including one near MAPRE2. Methods: We generated a loss-of-function mapre2 zebrafish mutant using CRISPR/Cas9. Embryonic hearts at 3 day-post-fertilization (dpf) were studied using voltage mapping whereas adult fish were studied using ECG and patch clamp. Results: In the KO vs. WT siblings, voltage mapping showed a decrease in ventricular conduction velocity (31.4±3.7 vs. 21.1±3.0 mm/s; P<0.05) and action potential (AP) upstroke velocity (Vmax; 98.6±2.5 vs. 81.2±7.6 1/s; P<0.05), as well as an increase in ventricular AP duration (APD) at 80% repolarization (260±16 vs. 346±29 ms; P<0.05). ECG of adult fish showed a decrease in HR (123±8 vs. 93±4 bpm; P<0.05), and an increase in QRS duration (33.3±2.1 vs. 42.9±1.6 ms; P<0.01) and QTc (352±18 vs. 437±17; P<0.05). Patch clamp of isolated ventricular myocytes showed a decrease in Vmax (165±12 vs. 91±10 1/s; P<0.0001), ~45% reduction in Na+ current density (P<0.05), and an increase in APD at 90% repolarization (170±9 vs. 219±11 ms; P<0.01). Significance: Genetic ablation of mapre2 led to a decrease in Na+ current function, a hallmark of BrS, as well as APD prolongation which may be related to an increase in late Na+ current and/or dysfunction in Ca2+ or K+ currents. MAPRE2 may be a novel causative gene in some BrS patients.
Precision Genomic Medicine in The Plain Communities and its Impact on The Plain and General Population
Lina Ghaloul-Gonzalez
Compare
Precision Genomic Medicine in The Plain Communities and its Impact on The Plain
and General Population Image

Connect with Lina @Ghaloul-Gonzalez-77 on Slack

Direct Message the author

Background: The Plain people (Amish and Mennonites) originated from founder populations with subsequent genetic bottlenecks and genetic drift; leading to loss of diversity and altered genetic disease burden. Western PA Plain people are among the least genetically characterized Plain communities in the US. Due to the loss of population genomic complexity, inbreeding, and sociologic isolation, many of the genetic disorders in the Plain people are inherited in autosomal recessive fashion due to homozygous mutations. This genetic makeup and the large families facilitate whole exome sequencing (WES) analysis to find candidate variants that can be causes of disease. Methods: Plain families with unexplained medical conditions are recruited into this study. WES is performed on genomic DNA and further analysis and functional studies will be performed. Results: Preliminary findings have allowed us to add to the list of known conditions in the Plain communities. These novel variants were found in genes known to be associated with a disease but not reported previously in the Plain people. - AK2 (c.622T>C;p.Ser208Pro) causing novel phenotype for reticular dysgenesis. - PNPT1 (c.1925_1927delTGG;p.Val642del) causing mitochondrial disorder. - LRBA (c.787 C>G;p.Leu263Val) causing severe juvenile rheumatoid arthritis. - GUCY2C (c.2381A>T;p.Asp794Val) causing chronic secretory diarrhea. Significance: A multitude of genetic diseases are characterized in the Plain population, and this study will identify additional diseases, enabling a community-centric personalized medicine approach to care based on individual genetic risk. In addition, knowledge of genetic disorders originally developed through study of the Plain Populations, can subsequently be applied in the general population.
Rescue of NF1 cryptic splice site in exon 13 mRNA with antisense morpholino treatment
Elias Awad
Compare
Rescue of NF1 cryptic splice site in exon 13 mRNA with antisense morpholino treatment  Image

Connect with Elias @Awad-85 on Slack

Direct Message the author

Neurofibromatosis type 1 (NF1) is an autosomal dominant genetic disease with almost 3000 different pathogenic mutations in the NF1 gene having been identified. Approximately 27% of these identified mutations cause splicing errors to occur within pre-mRNA. c.1466A>C is a recurrent mutation in exon 13 of the NF1 gene that leads to creation of a cryptic splice site and aberrant splicing causing a 62 base pair deletion from the mRNA resulting in a frameshift. In efforts to model this mutation we have developed both a human cell line and mouse model. We developed an iPS cell line using CRISPR/Cas9 gene editing that is homozygous for the mutation. In addition we have created a mouse line with a humanized exon 13 containing the C.1466A>C variant. We are in the process of phenotyping this mouse line. In efforts to evaluate therapeutics, we created a Taqman assay to distinguish between wild type and mutant alleles.. We then sought to test an antisense morpholino designed to mask the cryptic splice site and restore normal splicing. Our initial testing indicates that the morpholino is able to repress the cryptic splice site and restore normal splicing.
A Formative Evaluation Plan for the Veterans Affairs Pharmacogenomic Testing for Veterans (PHASER) Program
Olivia Dong
Compare
A Formative Evaluation Plan for the Veterans Affairs Pharmacogenomic Testing for Veterans (PHASER) Program Image

Connect with Olivia @Dong-99 on Slack

Direct Message the author

BACKGROUND: In 2019, the Veterans Health Administration initiated Pharmacogenomics Testing for Veterans (PHASER), a clinical program that offers free, preemptive pharmacogenetic testing of 11 genes to up to 250,000 Veterans nationwide over 4 years. The pharmacogenetic test results can help providers optimize prescribing of nearly 30 drugs per Clinical Pharmacogenetic Implementation Consortium guidelines, with the long-term goal of improving drug efficacy and reducing side effects. In 2019, 9 VA sites implemented PHASER where 85 providers ordered pharmacogenomic testing in 501 patients. Our objective was to develop a formative evaluation plan to improve implementation efforts and estimate implementation costs. METHODS: For six months, two evaluators observed weekly, in-person national coordinating program operations meetings, where the team discussed program developments and implementation progress. They identified key topics for further inquiry via a structured survey. They then developed formative evaluation surveys rooted in implementation science frameworks (the Consolidated Framework for Implementation Research and the Theory of Diffusion of Innovations) to collect data from participating sites. RESULTS: Formative evaluation surveys will capture data on program materials, resource utilization, and implementation processes using questions adapted from published resources for two implementation frameworks. The Qualtrics web-based surveys (Provo, UT) will be distributed to personnel involved with PHASER implementation at current and future participating sites (i.e. site champions and coordinators, providers, laboratory and IT staff) during program set-up and Veteran enrollment phases. SIGNIFICANCE: Data from the formative evaluation surveys will help guide continuous program revisions to improve the implementation of the PHASER program.
Deep Mutational Scanning as a Platform for Determining Pathogenicity of Variants of Uncertain Significance in Human Transcription Factors
Dustin Baldridge
Compare
Deep Mutational Scanning as a Platform for Determining Pathogenicity of Variants of Uncertain Significance in Human Transcription Factors Image

Connect with Dustin @Baldridge-100 on Slack

Direct Message the author

The explosion of sequencing of patients with suspected monogenic disorders has resulted in many molecular diagnoses, although the expected diagnostic yield is limited to 20-40%, depending on the phenotypic features of the patients tested. One main reason for this major limitation is the current inability to correctly classify Variants of Uncertain Significance (VUS), predominantly due to an absence of functional assessments of an individual variant’s effects. A scalable solution to this critical problem in genomic medicine is the application of Deep Mutational Scanning (DMS), which involves the simultaneous assessment of tens of thousands of variants in a protein of interest, using a reliable, carefully calibrated, cell-based assay. We chose to apply this approach to clinically relevant transcription factors because of their technical feasibility for establishing high throughput cell-based assays. We ranked all 1,600 human transcription factors by their clinical relevance, using the number of alleles in ClinVar as a proxy for incidence, and intersected this list with those genes for which molecular assays are readily available, in order to identify the top candidates for investigation. As an example, we are establishing a DMS for the transcription factor, GLI3, using fluorescence-activated cell sorting (FACS) coupled to sequencing. This system provides a molecular read-out for clinically significant variants that disrupt sonic hedgehog (SHH) signaling. The end result is a “look-up table” that gives clinicians high confidence classification of all possible coding variants in a given transcription factor of interest.
Identifying Undiagnosed Alpha-1 Antitrypsin Deficiency and Environmental Disease Modifiers via Phenotype Risk Score
Bryce Schuler
Compare
Identifying Undiagnosed Alpha-1 Antitrypsin Deficiency and Environmental Disease Modifiers via Phenotype Risk Score Image

Connect with Bryce @Schuler-107 on Slack

Direct Message the author

BACKGROUND: Alpha-1 antitrypsin deficiency (AATD) is a genetic disorder that can cause liver disease often requiring transplantation. Many symptomatic individuals remain undiagnosed possibly due to differences in environment, genotype (“M” and “Z” denote the reference and most common pathogenic allele, respectively), and phenotype. We hypothesized that a weighted score of AATD features (phenotype risk score, PheRS) would identify individuals likely to have AATD. We also hypothesized that viral hepatitis contributes to AATD phenotypic heterogeneity and investigated hepatitis C with respect to liver transplantation. METHODS: Our population consisted of over 72,000 genotyped individuals linked to deidentified electronic health records. We created a PheRS of AATD features, assessed association with AATD genotypes, and tested for interactions of genotypes and hepatitis C infection status for the outcome of liver transplantation. Linear and logistic regressions were adjusted for age, sex, and principal components. RESULTS: PheRS was higher for individuals with a clinical AATD diagnosis when compared to those without (mean diagnosed vs. undiagnosed PheRS: ZZ=3.6 vs. 0.88; p=5.0e-9; MZ=3.2 vs. 0.96; p=3.1e-2). Among the 36 individuals with the highest PheRS, two had an AATD clinical diagnosis. There was a significant interaction between genotype and hepatitis C with liver transplantation (ZZ: odds ratio=1.23, p<1.4e-4; MZ: odds ratio=1.11, p=5.2e-13). SIGNIFICANCE: PheRS effectively predicted the presence of an AATD diagnosis and therefore identified an enriched subset for subsequent AATD screening. Simultaneous presence of AATD and hepatitis C increases risk of liver transplantation notably among MZ heterozygotes, a genetic group typically ascribed low risk for liver failure.
Single-Cell Transcriptomic Profiling of Pediatric Ependymal Tumors in the Posterior Fossa Reveals Insights about their Oncogenic Pathways and Cell-of-Origin
Rachael Aubin
Compare
Single-Cell Transcriptomic Profiling of Pediatric Ependymal Tumors in the Posterior Fossa Reveals Insights about their Oncogenic Pathways and Cell-of-Origin Image

Connect with Rachael @Aubin-115 on Slack

Direct Message the author

BACKGROUND: Childhood ependymoma is a devastating type of brain cancer marked by its relapsing pattern and lack of effective chemotherapies. This shortage of treatment is partially due to limited understanding about the tumorigenesis and progression of these tumors. To remedy this lack of knowledge, we focused on studying the cell-of-origin and oncogenic pathways underlying the evolution of this disease. METHODS: Flash-frozen archived specimens and formalin-fixed paraffin-embedded slides were acquired from the Children’s Brain Tumor Tissue Consortium (CBTTC). Since conventional single-cell RNA-seq methods require fresh tissue, we employed a recently developed single-nuclei RNA-seq method (known as sNucDrop-seq). Utilizing manifold-learning, clustering, and graph-based methods, we reveal the cellular composition, developmental lineages, molecular pathways, and gene expression signatures of these tumors. Additionally, we combined our single-nuclei gene signatures with existing bulk transcriptomic data to deconvolve cell type abundance of tumors profiled with bulk RNA-seq. RESULTS: We have profiled over 13,000 cells from 5 primary posterior fossa tumors. Upon clustering cells and performing differential expression analysis, we identified 8 cell populations many of which resemble known CNS cell types. Utilizing graph-based techniques, we identified gene expression patterns among the malignant cells that suggest an amplification of canonical beta-catenin-dependent WNT signaling. Finally, we demonstrate that posterior fossa tumors can be stratified into two subtypes based on their transcriptome. These subtypes are associated with differences in vascularization, age of diagnosis, and methylation profiles. SIGNIFICANCE: This research will substantially advance our molecular understanding of tumorigenesis, progression, and relapse for childhood ependymoma and motivate future functional studies.
Rare pathogenic germline variants in sporadic neuroblastoma are nearly universally inherited.
Emily Blauel
Compare
Rare pathogenic germline variants in sporadic neuroblastoma are nearly universally inherited. Image

Connect with Emily @Blauel-156 on Slack

Direct Message the author

Background: The frequency of germline mutations in cancer predisposition genes (CPGs) is higher than previously estimated, with recent studies showing pathogenic mutations in up to 10% of childhood cancer patients. To date, no studies have investigated heritability of these variants. With only 1-2% of neuroblastoma cases presenting with a family history of disease (in the setting of neuroblastoma predisposition genes ALK and PHOX2B), the vast majority appear to arise sporadically. Methods: Through the Gabriella Miller Kids First program, we performed whole genome sequencing on child-parent triads (n = 493) and dyads (n = 100), along with whole genome and exome sequencing on matched tumor DNA (n = 366) and RNA-sequencing on matched tumor RNA (n = 228). Pathogenic and likely pathogenic (P/LP) germline variants in 198 pre-determined CPGs were identified and assessed for inheritance. Tumor sequencing was analyzed for relevance to tumorigenesis. Results: We observed 67 P/LP germline variants in CPGs across 64 probands, or 10.8% of our cohort. Of the 59 probands with trio data, 97% of the P/LP variants were inherited (47% maternal, 53% paternal). Several CPGs showed enrichment of P/LP variants. We observed one canonical ALK mutation, but no PHOX2B mutations. Significance: In our large cohort, we showed that approximately 1 in 10 patients with neuroblastoma harbor a rare P/LP germline mutation in a CPG. Importantly, we have shown that the majority of these variants are inherited. Work is ongoing to understand differences in penetrance between probands and parents, and each variant’s role in tumorigenesis.
Geographically and socioeconomically disadvantaged patients use a higher proportion of drugs with potentially actionable pharmacogenetic interactions
Rachel Dalton
Compare
Geographically and socioeconomically disadvantaged patients use a higher proportion of drugs with potentially actionable pharmacogenetic interactions Image

Connect with Rachel @Dalton-160 on Slack

Direct Message the author

Background: Pharmacogenetic testing can aid in optimizing therapy of drugs with Clinical Pharmacogenomics Implementation Consortium (CPIC) guidelines. We hypothesized that CPIC guideline drugs are used at higher rates in underserved populations because many are available as low-cost generics. This study compared CPIC drug usage with geographic and socioeconomic indices in University of Florida Health (UF-Health) patients. Methods: Electronic health record data from 2017 (demographics, drug lists, home zip code tabulation areas (ZCTAs)) were collected for adult patients. Geographic access scores (AS; range 1-100, 100 is greatest access) from ZCTA centroids to primary care providers were derived using the two-step floating catchment area method. US Census Bureau data were used for ZCTA-level socioeconomic indices. A negative binomial mixed-effects model was used to assess the relationship between CPIC prescription count and age, sex, race, ethnicity, total prescriptions, census data, and AS with ZCTA as a random effect. Results: 58,428 patients were analyzed. Patients with lower AS had more CPIC prescriptions; every 10 unit decrease in AS was associated with increases of 2.1% in CPIC prescriptions overall (p<0.001) and 4.6% in African Americans (p<0.05). Patient age (p<0.001), total prescription count (p<0.001), African American race (p=0.01), and residence in ZCTAs with more households below the federal poverty level (p<0.001) were associated with increased CPIC prescriptions. Significance: Underserved patients may further benefit from pharmacogenetic testing to maximize the utility of prescriber encounters and ensure safe and efficacious use of CPIC drugs. Further research is needed to confirm the generalizability of these findings.
CaBagE: a Cas9-based background elimination strategy for targeted, long-read DNA sequencing
Amelia Wallace
Compare
CaBagE: a Cas9-based background elimination strategy for targeted, long-read DNA sequencing Image

Connect with Amelia @Wallace-168 on Slack

Direct Message the author

BACKGROUND: Significant portions of the human genome are difficult to interrogate with short-read sequencing due to paralogy, complex haplotype structures, and tandem repeats. In total, over 700 genes and 60 megabases of the non-coding genome, which comprise numerous disease-associated loci, remain inaccessible to short reads. Long-read sequencing technologies, such as Oxford Nanopore’s MinION, enable direct measurement of complex loci without introducing many of the biases inherent to short-read methods, though they suffer from relatively lower throughput. This has motivated recent efforts to develop amplification-free strategies to target and enrich loci of interest for subsequent sequencing with long reads. Here, we present CaBagE, a novel method for target enrichment that is efficient and uniquely useful for sequencing large, structurally complex targets without requiring large, highly intact DNA molecules as input. METHODS: CaBagE enables targeted sequencing of structurally complex, medically-relevant loci by combining Cas9 and amplification-free sequence capture methods with Nanopore long-read sequencing technology. RESULTS: We demonstrate the utility of CaBagE in 2 ALS patients with C9orf72 repeat expansions and in 1 healthy carrier of structural variants in BRCA1 to show its generalizability as a genomic medicine tool that is fast, inexpensive, and useful in DNA samples of varying quality. SIGNIFICANCE: CaBagE is a rapid, adaptable method that can be used to illuminate structurally complex regions of the ‘dark genome’ that may underlie human disease. Although other target enrichment strategies require extremely high molecular weight DNA , CaBagE is amenable to shorter input fragments and can therefore be used with biobanked samples.
Finding order in disorder: quantifying genetic variability associated with protein features
Jason Miller
Compare
Finding order in disorder: quantifying genetic variability associated with protein features Image

Connect with Jason @Miller -187 on Slack

Direct Message the author

Background: Advances in sequencing and genotyping of individuals has led to a wealth of data that is being used to uncover the genetic underpinnings of complex traits and diseases. Although in-silico methods (i.e. CADD) have been developed to predict the impact of a change at the DNA level, they often have much poorer accuracy in regions that encode disordered portions of proteins. While intrinsically disordered regions (IDRs) tend to evolve more rapidly than more structured regions of proteins, there is evidence of purifying selection associated with functional regions within the IDRs such as post-translational modifications. We hypothesize identification of these functional regions under selection could be used in future methods for improved performance of pathogenic SNV detection. Methods: We designed a workflow to measure the frequency of common and rare variants in different protein domains, including IDRs using data from the 1000 genomes project. Results: Significant differences (p-value < 1x10^-4) were identified between common and rare variants among disordered and helix regions and across the entire proteome. As a next step, specific functional regions in IDRs will be examined. Significance: This approach was applied across individuals of diverse ancestral backgrounds as a proof of concept study for later testing in the context of specific diseases. This study highlights the importance of including protein structure information when studying genomics.
SPLiT-Seq Demultiplexing: an analytical pipeline for iteratively barcoded single-cell RNA-Seq
Paul Ranum
Compare
SPLiT-Seq Demultiplexing: an analytical pipeline for iteratively barcoded single-cell RNA-Seq 
 Image

Connect with Paul @Ranum-154 on Slack

Direct Message the author

Background: Single-cell RNA-Seq (scRNA-Seq) studies of the mouse and human brain have revealed novel cell types and anatomical features in the basal ganglia, a region that rapidly degenerates in Huntington’s disease. Investigating neurodegeneration in brain regions with small substructures will require fine spatial and temporal resolution not possible with popular droplet-based scRNA-Seq methodologies. To gain high spatiotemporal resolution at low cost we have implemented SPLiT-Seq a single-cell RNA-Seq methodology that utilizes iterative barcoding to encode sample and single-cell identities. Here, we report “SPLiT-Seq Demultiplexing” our analytical pipeline for processing iteratively barcoded single-cell data. Methods: SPLiT-Seq Demultiplexing performs end to end single-cell read assignment taking a fastq file as input and outputting a genes / cells counts matrix. Single-cell read assignment accuracy was evaluated using a computationally generated dataset of barcoded reads at maximum theoretical diversity in known quantities. Results: SPLiT-Seq Demultiplexing is the primary tool currently available for processing split-pool barcoded scRNA-Seq datasets. It obtains single-cell assignment accuracy of 100% at an error threshold of 1. Speed optimizations have been implemented facilitating demultiplexing of datasets > 100 million reads. The resulting analysis pipeline informed several methodological changes to the published SPLiT-Seq library preparation to reduce concatemeric product formation and improve the percentage of reads assigned to genes. Significance: SPLiT-Seq Demultiplexing is a fast and accurate open source tool available for processing iteratively barcoded scRNA-Seq data. It represents a valuable bioinformatic resource and facilitates scRNA-Seq at the fine spatiotemporal resolution needed to interrogate neurodegeneration in the mouse and human brain.
A patient’s own breast tumor provides evidence for improved treatment decisions. Spoiler... she's still alive!
Matthew Bailey
Compare
A patient’s own breast tumor provides evidence for improved treatment decisions. Spoiler... she

Connect with Matthew @Bailey-141 on Slack

Direct Message the author

BACKGROUND: With more than 430 medications approved to prevent or treat cancer (n=69 compounds for breast cancer), how do clinicians know which medicine to prescribe? Our lab recently developed protocols for integrating drug screening and genomic vulnerabilities from patient-derived tumor models. By combining these data, we can provide clinicians with answers to the following questions: what drug worked and why a drug worked. METHODS: We collected genomic and drug screening data from organoids obtained from both living and deceased patients with advanced breast cancer. Each patient tumor underwent comprehensive molecular characterization using next-generation DNA and RNA sequencing. Finally, we screened more than 40 compounds (on-label and off-label) on each patient-derived organoid. RESULTS: First, patient-derived organoids maintain molecular characteristics of the primary tumor, i.e., driver mutations, transcriptomic profiles, and copy number variations. Second, differences in drug responses within a single patient provide information necessary for drug prioritization. Fortunately, with IRB approval, our group recently returned our results to a treating physician, with an FDA approved compound that displayed the strong cytotoxic response in their patient's tumor-organoid. To date, this informed treatment decision has outperformed the patient's previous 'uninformed' treatments. Third, by grouping responders and non-responders, we have identified novel biomarkers that associate with drug efficacy. SIGNIFICANCE: Unlike many approaches that rely on genomic prediction alone, our unique functional approach delivers robust forecasts that exceed current limitations of drug-gene databases. Additionally, as we expand our database of screened tumor-organoids, we expect to identify new biomarkers that will guide the drug selection process.
Exploring the role of TDP-43 in transposable element regulation
Rosa Ma
Compare
Exploring the role of TDP-43 in transposable element regulation Image

Connect with Rosa @Ma-49 on Slack

Direct Message the author

BACKGROUND: Transposable elements (TEs) are repetitive mobile DNA sequences that are usually repressed by histone H3K9 trimethylation (H3K9me3). De-repression of TEs may be part of the ALS disease process. In 97% of ALS patients, TDP-43, a nucleic acid binding protein, is mislocalized from the nucleus to the cytoplasm. Recent studies have shown that nuclear depletion of TDP-43 is associated with increased chromatin accessibility to TEs and an up-regulation of TE transcripts. Since TDP-43 is enriched in H3K9me3-marked heterochromatic regions and interacts with proteins essential for heterochromatin formation, TDP-43 may repress TEs through modifying H3K9me3. METHODS: To test this hypothesis, HAP-1 cells will be transduced with a lentiviral reporter construct expressing mCherry. Cells that have reporters integrated into a repressive chromatin environment will exhibit lower mCherry expression and will be isolated by FACS. CRISPR/Cas9 will then be used to knock out TDP-43 in these cells, and another round of FACS will be used to enrich for cells expressing high levels of mCherry. Linear amplification-mediated PCR followed by read mapping will be used to identify whether the reporters are inserted in sites normally marked by H3K9me3. RESULTS: It is predicted that the reporters integrated at H3K9me3 marked region will be de-repressed upon TDP-43 depletion. SIGNIFICANCE: The results from this study will provide mechanistic support for the observation that transcription of TEs is dysregulated in cells with TDP-43 depletion and will enrich our current understanding of the toxicity mechanism driven by the loss of nuclear TDP-43.
Finding the missing heritability in epilepsy using haplotypes
Pouya Khankhanian
Compare
Finding the missing heritability in epilepsy using haplotypes Image

Connect with Pouya @Khankhanian-105 on Slack

Direct Message the author

BACKGROUND: Epilepsy is a debilitating chronic neurologic illness affecting 60 million people worldwide, it costs the U.S. health system $3.4 billion annually. It is clear from family pedigrees that patients have a genetic etiology. However, the genetic cause can only be found for ~1/3 of patients (based on standard whole exome sequencing, copy number variation, insertion-deletion detection, and GWAS). The etiology of the genetic epilepsy for the remaining ~2/3 of patients is termed the “missing heritability”. METHODS: A secondary analysis of GWAS data leveraging haplotypes to find the missing heritability. GWAS data come from the largest international consortium of epilepsy patients (Epi25). Haplotypes are estimated by standard genome-wide haplotyping software (EAGLE) supplemented by internally developed EM algorithms. Rare haplotypes are assessed for disease risk by evaluating level of conservation, comparison to biobanks, assessment of functionality, and association with known offending gene classes. Common haplotypes are assessed for disease risk by standard GWAS techniques of logistic and linear regression. RESULTS: We hope to find a likely-causative genetic variant for a larger percentage of patients than previous methods (~33%). We also hope to leverage haplotypes into a more robust polygenic risk score (PRS) explain a larger fraction of risk across the population (Nagelkerke-R2) than has been shown with traditional PRS (~0.04-0.08). SIGNIFICANCE: The value of finding the genetic etiology for a given patient is proven dramatic decreased cost on the health system, as well as the opportunity for targeted therapy to improve the outcome for the patient.
200 products«1 of 10»
200 products«1 of 10»
Genomic Medicine
Compare
Exome-by-phenome-wide association with electronic health record phenotypes Image
Joseph Park
Compare
MAPRE2 may be a novel Brugada syndrome gene Image
MAPRE2 may be a novel Brugada syndrome gene
Connect with David @Chiang-63 on Slack
David Chiang
Compare
Precision Genomic Medicine in The Plain Communities and its Impact on The Plain
and General Population Image
Lina Ghaloul-Gonzalez
Compare
Exploring the role of TDP-43 in transposable element regulation Image
Rosa Ma
Compare
Finding the missing heritability in epilepsy using haplotypes Image
Finding the missing heritability in epilepsy using haplotypes
Connect with Pouya @Khankhanian-105 on Slack
Pouya Khankhanian
200 products«1 of 10»