Biology subjects

Patel, C. J.

Publications and source records attributed to Patel, C. J..

6 recordsLinked to original sources

bioRxiv2018

A transcript-wide association study in physical activity intervention implicates molecular pathways in chronic disease

BackgroundPhysical activity is associated with decreased risk for several chronic and acute conditions including obesity, diabetes, cardiovascular disease, mental health and aging. However, the biological mechanisms associated with this decreased risk are elusive. One way to ascertain biological changes influenced by physical activity is by monitoring changes in how genes are expressed. In this investigation, we conducted a transcriptome-wide association study of physical activity, meta-analyzing 20 independent studies to increase power for discovery of genes expressed before and after physical activity. Further, we hypothesize that genes identified in physical activity are expressed in obesity, inflammation, major depressive disorder and healthy aging.\n\nResultsOur analysis identified thirty (30) transcripts induced by physical activity (PA signature), at an FDR < 0.05. Twenty (20) of these transcripts, including COL4A3, CAMKD1, SLC4A5, EPS15L1, RBM33, and CACNG1, are up-regulated and ten (10) transcripts including CRY1, ZNF346, SDF4, ANXA1 and YWHAZ are down-regulated. We find that several of these physical activity transcripts are associated and biologically concordant in direction with body mass index, white blood cell count, and healthy aging.\n\nConclusionspowerful approach, we found thirty genes that were putatively influenced by physical activity, eight of which are inversely associated with body mass index, thirteen inversely associated with white blood cell count, and three associated and concordant with healthy aging. One gene was significant and concordant with major depressive disorder. These results highlight the potential molecular basis for the protective benefit of physical activity for a broad set of chronic conditions.

genomics↗

bioRxiv2017

Toward Capturing the Exposome: Exposure Biomarker Variability and Co-Exposure Patterns in the Shared Environment

BACKGROUNDAlong with time, variation in the exposome is dependent on the location and sex of study participants. One specific factor that may influence exposure co-variations is a shared household environment.\n\nOBJECTIVESTo examine the influence of shared household and partners sex in relation to the variation in 128 endocrine disrupting chemical (EDC) exposures among couples.\n\nMETHODSIn a cohort comprising 501 couples trying for pregnancy, we measured 128 (13 chemical classes) persistent and non-persistent EDCs and estimated 1) sex-specific differences; 2) variance explained by shared household; and 3) Spearmans rank correlation coefficients (rs) for females, males, and couples exposures.\n\nRESULTSSex was correlated with 8 EDCs including polyfluoroalkyl substances (PFASs) (p < 0.05). Shared household explained 43% and 41% of the total variance for PFASs and blood metals, respectively, but less than 20% for the remaining 11 EDC classes. Co-exposure patterns of the exposome were similar between females and males, with within-class rs higher for persistent and lower for non-persistent chemicals. Median rss of polybrominated compounds and urine metalloids were 0.45 and 0.09, respectively, for females (0.41 and 0.08 for males), whereas lower rss for these 2 classes were found for couples (0.21 and 0.04).\n\nCONCLUSIONSOverall, sex did not significantly affect EDC levels in couples. Individual, rather than shared environment, could be a major factor influencing the co-variation of 128 markers of the exposome. Correlations between exposures are lower in couples than in individual partners and have important analytical and sampling implications for epidemiological study.

epidemiology↗

bioRxiv2017

Aether: Leveraging Linear Programming For Optimal Cloud Computing In Genomics

Across biology we are seeing rapid developments in scale of data production without a corresponding increase in data analysis capabilities. Here, we present Aether (http://aether.kosticlab.org), an intuitive, easy-to-use, cost-effective, and scalable framework that uses linear programming (LP) to optimally bid on and deploy combinations of underutilized cloud computing resources. Our approach simultaneously minimizes the cost of data analysis while maximizing its efficiency and speed. As a test, we used Aether to de novo assemble 1572 metagenomic samples, a task it completed in merely 13 hours with cost savings of approximately 80% relative to comparable methods.

bioinformatics↗

bioRxiv2017

Combining Ensemble Learning Techniques and G-Computation to Investigate Chemical Mixtures in Environmental Epidemiology Studies

BackgroundAlthough biomonitoring studies demonstrate that the general population experiences exposure to multiple chemicals, most environmental epidemiology studies consider each chemical separately when assessing adverse effects of environmental exposures. Hence, the critical need for novel approaches to handle multiple correlated exposures.\n\nMethodsWe propose a novel approach using the G-formula, a maximum likelihood-based substitution estimator, combined with an ensemble learning technique (i.e. SuperLearner) to infer causal effect estimates for a multi-pollutant mixture. We simulated four continuous outcomes from real data on 5 correlated exposures under four exposure-response relationships with increasing complexity and 500 replications. The first simulated exposure-response was generated as a linear function depending on two exposures; the second was based on a univariate nonlinear exposure-response relationship; the third was generated as a linear exposure-response relationship depending on two exposures and their interaction; the fourth simulation was based on a non-linear exposure-response relationship with an effect modification by sex and a linear relationship with a second exposure. We assessed the method based on its predictive performance (Minimum Square error [MSE]), its ability to detect the true predictors and interactions (i.e. false discovery proportion, sensitivity), and its bias. We compared the method with generalized linear and additive models, elastic net, random forests, and Extreme gradient boosting. Finally, we reconstructed the exposure-response relationships and developed a toolbox for interactions visualization using individual conditional expectations.\n\nResultsThe proposed method yielded the best average MSE across all the scenarios, and was therefore able to adapt to the true underlying structure of the data. The method succeeded to detect the true predictors and interactions, and was less biased in all the scenarios. Finally, we could correctly reconstruct the exposure-response relationships in all the simulations.\n\nConclusionsThis is the first approach combining ensemble learning techniques and causal inference to unravel the effects of chemical mixtures and their interactions in epidemiological studies. Additional developments including high dimensional exposure data, and testing for detection of low to moderate associations will be carried out in future developments.

epidemiology↗

bioRxiv2017

Epigenome-Based Drug Repositioning in Acute Myeloid Leukemia

BackgroundRepositioning approved drugs for the treatment of new indications is a promising avenue to reduce the burden of drug development. Most currently available computational methods based on molecular evidence can only utilize gene expression for repositioning despite a growing interest in the epigenome in human disease. We recently described a novel repositioning method, ksRepo, that enables investigators to move beyond microarray-based gene expression and utilize a variety of other sources of molecular evidence, such as DNA methylation differences.\n\nMethodsWe downloaded differential DNA methylation data from two publicly available acute myeloid leukemia (AML) datasets, a cancer with known, extensive epigenomic perturbations. We consolidated CpGs-level to non-directional gene-level differential methylation using Browns correction to Fishers method. We then used ksRepo, which ignores directionality in disease- and gene-drug associations, to mine the resulting prioritized gene lists and and the Comparative Toxicogenomics Database (CTD) for predicted repositioning candidates.\n\nResultsWe successfully recovered four compounds that were significant (FDR < 0.05) in two AML datasets: cytarabine, alitretinoin, panobinostat, and progesterone. Cytarabine is the most commonly used frontline therapy for AML and alitretinoin, panobinostat, and progesterone have all been investigated for the treatment of AML.\n\nConclusions. Combining a method for consolidating CpG methylation to the gene level with ksRepo provides a pipeline for deriving drug repositioning hypotheses from differential DNA methylation. We claim that our platform can be extended to other diseases with epigenetic perturbations and to other epigenomic modalities, such as ChIP-seq.

bioinformatics↗

bioRxiv2017

Leveraging population-based clinical quantitative phenotyping for drug repositioning

Computational drug repositioning methods can scalably nominate approved drugs for new diseases, with reduced risk of unforeseen side effects. The majority of methods eschew individual-level phenotypes despite the promise of biomarker-driven repositioning. In this study, we propose a framework for discovering serendipitous interactions between drugs and routine clinical phenotypes in cross-sectional observational studies. Key to our strategy is the use of a healthy and non-diabetic population derived from the National Health and Nutrition Examination Survey, mitigating risk for confounding by indication. We combine complementary diagnostic phenotypes (fasting glucose and glucose response) and associate them with prescription drug usage. We then sought confirmation of phenotype-drug associations in un-identifiable member claims data from Aetna using a retrospective self-controlled case analysis approach. We identify bupropion hydrochloride as a plausible antidiabetic agent, suggesting that surveying otherwise healthy individuals cross-sectional studies can discover new drug repositioning hypotheses that have applicability to longitudinal clinical practice.

bioinformatics↗