Larson, Nicholas B.

AbstractVariance component models have gained popularity for genetic analyses, driven by their flexibility to simultaneously analyze multiple genetic variants in a gene by kernel statistics, and their ability to account for population stratification via genomic relationship matrices. For exploratory analyses with modest sample sizes and a potentially large number of variance components, it can be challenging to use standard maximum‐likelihood or restricted maximum‐likelihood methods to estimate variance components, because these iterative methods often fail to converge when likelihood surfaces are fairly flat, and standard‐likelihood ratio statistical tests are not adequate. To overcome these limitations, we developed a penalized‐likelihood model, whereby the penalty function follows the popular elastic‐net approach, applying both L1 and L2 penalties to the variance components. By simulations, we demonstrate the potential gain in power by using both L1 and L2 penalties, and results from our simulations suggest that assigning 80% of the penalty parameter to the L1 penalty and 20% to the L2 penalty provides a reasonable balance between false‐positive and false‐negative results. Larger sample size improves the properties of our methods, at the cost of longer computation time. Application of our methods to a study of the influence of DNA methylation on levels of cortisol in reaction to stress testing shows how our method can be used to prioritize findings for further functional studies.

e13535 Background: Pediatric patients (pts) receiving long-term vincristine treatment for acute lymphoblastic leukemia (ALL) who carry two variant alleles of CEP72 rs924607 experience more severe peripheral neuropathy (PN) at lower cumulative vincristine doses. The objective of this retrospective pharmacogenetic analysis was to determine if pts homozygous for rs924607 have increased PN risk when receiving short-term vincristine treatment for diffuse large B-cell lymphoma (DLBCL). Methods: Caucasian pts with DLBCL randomized to either of two vincristine-containing regimens (R-CHOP or DA-EPOCH-R) on the CALGB 50303 (Alliance) trial were genotyped for germline rs924607 via TaqMan. Vincristine was administered intravenously 1.4 mg/m2 (2 mg cap) on day 1 of each R-CHOP cycle (1.4 mg/m2/cycle) or 0.4 mg/m2/day (no cap) on days 1-4 of each DA-EPOCH-R cycle (1.6 mg/m2/cycle). Dosing was repeated every 21 days for 6 treatment cycles, for a maximum cumulative vincristine dose of 8.4 (R-CHOP) or 9.6 (DA-EPOCH-R) mg/m2 over 18 weeks. The primary analysis compared risk of NCI CTCAE grade 3 or higher PN in rs924607 homozygous variant pts versus wild-type allele carriers using multivariable Cox proportional hazards regression adjusted for treatment received (R-CHOP vs. DA-EPOCH-R), age (continuous), and sex (male vs. female). A secondary analysis was conducted only in pts assigned to the DA-EPOCH-R arm. Results: Of 307 pts included in the analysis, 32 (10.4%) experienced grade 3+ PN. PN occurrence was nominally lower in homozygous pts (5/59 = 8.5%) than wild-type carriers (27/248 = 10.9%). In the primary analysis, homozygous pts did not experience PN earlier in treatment (Hazard ratio (HR) = 1.04, (95% confidence interval: 0.39-2.75), one-sided p-value = 0.47). There was no significant difference between groups in dose-at-PN occurrence (homozygous (n = 5): 8.81 mg/m2 vs. wild-type carrier (n = 27):6.74 mg/m2, p > 0.05). The secondary analysis in pts on the DA-EPOCH-R arm (n = 147), in which the majority of PN events occurred (n = 27/32 = 84%), homozygous pts (n = 26) did not have significantly increased PN risk (HR = 1.30, p = 0.56). Conclusions: This retrospective analysis of the prospective CALGB 50303 trial did not find that pts homozygous for CEP72 rs924607 had increased PN risk during short-term vincristine treatment for DLBCL. Support: U10CA180821, U10CA180882, U24CA196171; https://acknowledgments.alliancefound.org . ClinicalTrials.gov Id: NCT00118209.

Background Opioid analgesics are among the most commonly prescribed medications in the United States. Opioids are useful in the management of pain, but their use can also adversely affect patients. However, studies reporting patterns of adverse events due to use of opioid analgesics in men and women are still scarce. Understanding the individual characteristics affecting response to opioids may allow clinicians to better individualize treatments. In this study, we investigated sex differences in the patterns of adverse events as a result of opioid use. Methods We used data from 8,457 patients participating in the Mayo Clinic Biobank “Right Drug, Right Dose, Right Time (RIGHT protocol)” study of pre‐emptive genotyping who were prescribed at least one of four opioid medications (codeine, tramadol, hydrocodone, or oxycodone) between 01/01/2004 and 12/31/2017. Adverse events due to use of these medications were collected from the allergy section in each patient’s electronical medical record, and were organized into four Medical Dictionary for Regulatory Activities (MedDRA) categories: gastrointestinal (e.g. nausea, vomiting), skin and subcutaneous tissue (e.g. rash, itching), psychiatric (e.g. anxiety, hallucinations), and nervous system (e.g. headache, dizziness). We modeled sex differences in the risk of adverse events by each opioid, and sex differences in the risk of each adverse event due to any opioid using logistic regression adjusted for age, race, ethnicity, and body mass index (BMI). Results Participants with at least one opioid prescription were 60.8% female, mostly white (94.2%) and non‐Hispanic (98.9%). Women (relative to men) were younger (median = 61.6 vs. 69.1, respectively), had lower BMI (median = 27.4 vs. 28.9), and were prescribed more codeine (18.8% vs. 15.5%). Overall, 456 (5.4%) persons had a note in the allergy section of their medical record indicating a problem with an opioid medication. Women were more likely than men to have a note in the allergy section referring to at least one opioid (6.6% and 3.5%, respectively; p < 0.001). In addition, among participants who had at least one note related to opioid use in their allergy section, women were more likely to have notes regarding gastrointestinal complications (OR [95% CI] = 2.07 [1.3 – 3.2], p = 0.002), but less likely to have notes regarding psychiatric issues (OR [95% CI] = 0.30 [0.2 – 0.5], p < 0.001). Conclusion Women were more likely than men to experience adverse events due to use of codeine, tramadol, hydrocodone, and oxycodone. In addition, the types of adverse reactions differed between men and women. These data suggest that further studies are needed to ensure that both men and women have optimal responses to opioid medications. Support or Funding Information This work was supported by the Mayo Clinic Specialized Center of Research Excellent on Sex Differences, the Mayo Clinic Center for Individualized Medicine, the Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, and the National Institutes of Health grants [U19 GM61388 (The Pharmacogenomics Research Network), R01 GM28157, U01 HG005137, R01 GM125633, R01 AG034676 (The Rochester Epidemiology Project)], and [U01 HG06379 and U01 HG06379 Supplement (The Electronic Medical Record and Genomics (eMERGE) Network)].

Higher levels of hepatocyte growth factor (HGF) have been associated with the presence of peripheral arterial disease (PAD), but prospective associations are unknown. We examined the association of circulating HGF levels with incident PAD. Between 2000 and 2002, HGF was measured in 6742 Multi-Ethnic Study of Atherosclerosis participants without PAD. Incident clinical PAD, adjudicated on the basis of a positive history for the presence of disease-related symptoms or treatment, was ascertained through 2015. Incident low ankle-brachial index (ABI), defined as an ABI < 0.9 and a decline of ≥ 0.15, was assessed among 5736 individuals who had an ABI > 0.9 at baseline and ≥1 follow-up ABI measurement 3 to 10 years later. There were 116 clinical PAD and 197 low ABI events that occurred over a median follow-up of 14 and 9 years, respectively. After adjustment for demographic and clinical variables, a standard deviation increment of HGF (303 ng/L) was associated with an increased risk of clinical PAD (hazard ratio: 1.21; 95% confidence interval [CI]: 1.05-1.39) but not a low ABI (rate ratio: 1.03; 95% CI: 0.85-1.25). In conclusion, higher HGF levels were modestly associated with an increased risk of developing clinical PAD.

Background: Current cardiovascular disease (CVD) risk scores are derived from research cohorts and are particularly inaccurate in women, older adults, and those with missing data. To overcome these limitations, we aimed to develop a cohort to capitalize on the depth and breadth of clinical data within electronic health record (EHR) systems in order to develop next-generation sex-specific risk prediction scores for incident CVD. Methods: All individuals 30 years of age or older residing in Olmsted County, Minnesota on 1/1/2006 were identified. We developed and validated algorithms to define a variety of risk factors, thus building a comprehensive risk profile for each patient. Outcomes including myocardial infarction (MI), percutaneous intervention (PCI), coronary artery bypass graft (CABG), and CVD death were ascertained through 9/30/2017. Results: We identified 73,069 individuals without CVD (Table). We retrieved a total of 14,962,762 lab results; 14,534,466 diagnoses; 17,062,601 services/procedures; 1,236,998 outpatient prescriptions; 1,079,065 heart rate measurements; and 1,320,115 blood pressure measurements. The median number of blood pressure and heart rate measurements ascertained per individuals were 11 and 9, respectively. The five most prevalent conditions were: hypertension, hyperlipidemia, arthritis, depression, and cardiac arrhythmias. During follow-up 1,455 MIs, 1,581 PCI, 652 CABG, and 2,161 CVD-related deaths occurred. Conclusions: We developed a cohort with comprehensive risk profiles and follow-up for each patient. Using sophisticated machine learning approaches, this electronic cohort will be utilized to develop next-generation sex-specific CVD risk prediction scores. These approaches will allow us to address several challenges with use of EHR data including the ability to 1) deal with missing values, 2) assess and utilize a large number of variables without over-fitting, 3) allow non-linear relationships, and 4) use time-to-event data.

Objective The aim of this study was to grade cartilage damage in a swine model of osteoarthritis using a whole-body photon-counting detector (PCD) CT. Materials and Methods A multienergy phantom containing gadolinium (Gd) (2, 4, 8, and 16 mg/mL) and hydroxyapatite (200 and 400 mg/cc) was scanned using a PCD-CT system (48 × 0.25 mm collimation, 80 kV, 800 mAs, D50 reconstruction kernel) to serve as calibration for material decomposition and to assess quantification accuracy. Osteoarthritis was induced in Yucatan miniswine (n = 8) using 1.2 mg monoiodoacetate (MIA) injected into a randomized knee, whereas the contralateral control knee received saline. Twenty-one days later, a contrast bolus (gadoterate meglumine, 4 mL/knee) was intra-articularly administered into both knees. The knees were simultaneously scanned on the PCD-CT system (48 × 0.25 mm collimation, 80 kV, 800 mAs). Multienergy images were reconstructed with a sharp “V71” kernel and a quantitative “D50” kernel. Image denoising was applied to the V71 images before grading cartilage damage, and an iterative material decomposition technique was applied to D50 images to generate the Gd maps. Two radiologists blinded to the knee injection status graded the cartilage integrity based on a modified International Cartilage Repair Society scoring system. Histology was performed on excised cartilage using methylene blue/basic fuchsin. Statistical analysis of grade distribution was performed using an exact test of omnibus symmetry with P < 0.05 considered significant. Results Material decomposed images from the multienergy phantom scan showed delineation and quantification of Gd and hydroxyapatite with a root-mean-squared error of 0.3 mg/mL and 18.4 mg/cc, respectively. In the animal cohort, the radiologists reported chondromalacia in the MIA knees with International Cartilage Repair Society scores ranging from grade 1 (cartilage heterogeneity, n = 4 knees) to grade 3 (up to 100% cartilage loss, n = 4 knees). Grade 1 was characterized by cartilage heterogeneity and increased joint space in the patellofemoral compartment, whereas grade 3 was characterized by cartilage erosion and bone-on-bone articulation in the patellofemoral compartment. All control knees were scored as grade 0 (normal cartilage). Significant difference (P = 0.004) was observed in the grade distribution between the MIA and control knees. Gross examination of the excised knees showed cartilage lesions in the grade 3 MIA knees. The Gd maps from material decomposition showed lower contrast levels in the joint space of the MIA knee compared with the contralateral control knee due to joint effusion. Histology revealed chondrocyte loss in the MIA knee cartilage confirming the chondrotoxic effects of MIA on cartilage matrix. Conclusions We demonstrated a high-resolution and quantitative PCD-CT arthrography technique for grading cartilage damage in a large animal model of osteoarthritis. Photon-counting detector CT offers simultaneous high-resolution and multienergy imaging capabilities that allowed morphological assessment of cartilage loss and quantification of contrast levels in the joint as a marker of joint disease. Cartilage damage in the MIA knees was graded using PCD-CT images, and the image-based findings were further confirmed using histology and gross examination of the excised knees.

Abstract Motivation DNA methylation can be measured at the single CpG level using sodium bisulfite conversion of genomic DNA followed by sequencing or array hybridization. Many analytic tools have been developed, yet there is still a high demand for a comprehensive and multifaceted tool suite to analyze, annotate, QC and visualize the DNA methylation data. Results We developed the CpGtools package to analyze DNA methylation data generated from bisulfite sequencing or Illumina methylation arrays. The CpGtools package consists of three types of modules: (i) ‘CpG position modules’ focus on analyzing the genomic positions of CpGs, including associating other genomic and epigenomic features to a given list of CpGs and generating the DNA motif logo enriched in the genomic contexts of a given list of CpGs; (ii) ‘CpG signal modules’ are designed to analyze DNA methylation values, such as performing the PCA or t-SNE analyses, using Bayesian Gaussian mixture modeling to classify CpG sites into fully methylated, partially methylated and unmethylated groups, profiling the average DNA methylation level over user-specified genomics regions and generating the bean/violin plots and (iii) ‘differential CpG analysis modules’ focus on identifying differentially methylated CpGs between groups using different statistical methods including Fisher’s Exact Test, Student’s t-test, ANOVA, non-parametric tests, linear regression, logistic regression, beta-binomial regression and Bayesian estimation. Availability and implementation CpGtools is written in Python under the open-source GPL license. The source code and documentation are freely available at https://github.com/liguowang/cpgtools. Supplementary information Supplementary data are available at Bioinformatics online.