Biology subjects

O'Donnell, C. J.

Publications and source records attributed to O'Donnell, C. J..

4 recordsLinked to original sources

bioRxiv2019

Genetic predisposition to myeloproliferative neoplasms implicates hematopoietic stem cell biology

Myeloproliferative neoplasms (MPNs) are blood cancers characterized by excessive production of mature myeloid cells that result from the acquisition of somatic driver mutations in hematopoietic stem cells (HSCs)1. While substantial progress has been made to define the causal somatic mutation profile for MPNs2, epidemiologic studies indicate a significant heritable component for the disease that is among the highest known for all cancers3. However, only a limited set of genetic risk loci have been identified, and the underlying biological mechanisms leading to MPN acquisition remain unexplained. Here, to define the inherited risk profile, we conducted the largest genome-wide association study of MPNs to date (978,913 individuals with 3,224 cases) and identified 14 genome-wide significant loci, as well as a polygenic signature that increases the odds for disease acquisition by nearly 3-fold between the top and median deciles. Interestingly, we find a shared genetic architecture between MPN risk and several hematopoietic traits spanning distinct lineages, as well as an association between increased MPN risk and longer leukocyte telomere length, collectively implicating HSC function and self-renewal. Strikingly, we find a significant enrichment for risk variants mapping to accessible chromatin in HSCs compared with other hematopoietic populations. Finally, gene mapping identifies modulators of HSC biology and targeted variant-to-function analyses suggest likely roles for CHEK2 and GFI1B in altering HSC function to confer disease risk. Overall, we demonstrate the power of human genetic studies to illuminate a previously unappreciated mechanism for MPN risk through modulation of HSC function.

genetics↗

bioRxiv2019

High-throughput Multimodal Automated Phenotyping (MAP) with Application to PheWAS

ObjectiveElectronic health records (EHR) linked with biorepositories are a powerful platform for translational studies. A major bottleneck exists in the ability to phenotype patients accurately and efficiently. The objective of this study was to develop an automated high-throughput phenotyping method integrating International Classification of Diseases (ICD) codes and narrative data extracted using natural language processing (NLP).\n\nMethodWe developed a mapping method for automatically identifying relevant ICD and NLP concepts for a specific phenotype leveraging the UMLS. Aggregated ICD and NLP counts along with healthcare utilization were jointly analyzed by fitting an ensemble of latent mixture models. The MAP algorithm yields a predicted probability of phenotype for each patient and a threshold for classifying subjects with phenotype yes/no. The algorithm was validated using labeled data for 16 phenotypes from a biorepository and further tested in an independent cohort PheWAS for two SNPs with known associations.\n\nResultsThe MAP algorithm achieved higher or similar AUC and F-scores compared to the ICD code across all 16 phenotypes. The features assembled via the automated approach had comparable accuracy to those assembled via manual curation (AUCMAP 0.943, AUCmanual 0.941). The PheWAS results suggest that the MAP approach detected previously validated associations with higher power when compared to the standard PheWAS method based on ICD codes.\n\nConclusionThe MAP approach increased the accuracy of phenotype definition while maintaining scalability, facilitating use in studies requiring large scale phenotyping, such as PheWAS.

bioinformatics↗

bioRxiv2017

A Comprehensive Evaluation of the Genetic Architecture of Sudden Cardiac Arrest

BackgroundSudden cardiac arrest (SCA) accounts for 10% of adult mortality in Western populations. While several risk factors are observationally associated with SCA, the genetic architecture of SCA in the general population remains unknown. Furthermore, understanding which risk factors are causal may help target prevention strategies.\n\nMethodsWe carried out a large genome-wide association study (GWAS) for SCA (n=3,939 cases, 25,989 non-cases) to examine common variation genome-wide and in candidate arrhythmia genes. We also exploited Mendelian randomization methods using cross-trait multi-variant genetic risk score associations (GRSA) to assess causal relationships of 18 risk factors with SCA.\n\nResultsNo variants were associated with SCA at genome-wide significance, nor were common variants in candidate arrhythmia genes associated with SCA at nominal significance. Using cross-trait GRSA, we established genetic correlation between SCA and (1) coronary artery disease (CAD) and traditional CAD risk factors (blood pressure, lipids, and diabetes), (2) height and BMI, and (3) electrical instability traits (QT and atrial fibrillation), suggesting etiologic roles for these traits in SCA risk.\n\nConclusionsOur findings show that a comprehensive approach to the genetic architecture of SCA can shed light on the determinants of a complex life-threatening condition with multiple influencing factors in the general population. The results of this genetic analysis, both positive and negative findings, have implications for evaluating the genetic architecture of patients with a family history of SCA, and for efforts to prevent SCA in highrisk populations and the general community.

genetics↗

bioRxiv2017

Machine learning identifies SNPs predictive of advanced coronary artery calcium in ClinSeq(R) and Framingham Heart Study cohorts

One goal of personalized medicine is leveraging the emerging tools of data science to guide medical decision-making. Achieving this using disparate data sources is most daunting for polygenic traits and requires systems level approaches. To this end, we employed random forests (RF) and neural networks (NN) for predictive modeling of coronary artery calcification (CAC), which is an intermediate end-phenotype of coronary artery disease (CAD). Model inputs were derived from advanced cases in the ClinSeq(R) discovery cohort (n=16) and the FHS replication cohort (n=36) from 89th-99th CAC score percentile range, and age-matching controls (ClinSeq(R) n=16, FHS n=36) with no detectable CAC (all subjects were Caucasian males). These inputs included clinical variables (CLIN), genotypes of 57 SNPs associated with CAC in past GWAS (SNP Set-1), and an alternative set of 56 SNPs (SNP Set-2) ranked highest in terms of their nominal correlation with advanced CAC state in the discovery cohort. Predictive performance was assessed by computing the areas under receiver operating characteristics curves (AUC). Within the discovery cohort, RF models generated AUC values of 0.69 with CLIN, 0.72 with SNP Set-1, and 0.77 with their combination. In the replication cohort, SNP Set-1 was again more predictive (AUC=0.78) than CLIN (AUC=0.61), but also more predictive than the combination (AUC=0.75). In contrast, in both cohorts, SNP Set-2 generated enhanced predictive performance with or without CLIN (AUC> 0.8). Using the 21 SNPs of SNP Set-2 that produced optimal predictive performance in both cohorts, we developed NN models trained with ClinSeq(R) data and tested with FHS data and replicated the high predictive accuracy (AUC>0.8) with several topologies, thereby identifying several potential susceptibility loci for advanced CAD. Several CAD-related biological processes were found to be enriched in the network of genes constructed from these loci. In both cohorts, SNP Set-1 derived from past CAC GWAS yielded lower performance than SNP Set-2 derived from \"extreme\" CAC cases within the discovery cohort. Machine learning tools hold promise for surpassing the capacity of conventional GWAS-based approaches for creating predictive models utilizing the complex interactions between disease predictors intrinsic to the pathogenesis of polygenic disorders.

genomics↗