A semi-automated workflow for adverse outcome pathway hypothesis generation: The use case of non-genotoxic induced hepatocellular carcinoma.

The utility of the Adverse Outcome Pathway (AOP) concept has been largely recognized by scientists, however, the AOP generation is still mainly done manually by screening through evidence and extracting probable associations. To accelerate this process and increase the reliability, we have developed an semi-automated workflow for AOP hypothesis generation. In brief, association mining methods were applied to high-throughput screening, gene expression, in vivo and disease data present in ToxCast and Comparative Toxicogenomics Database. This was supplemented by pathway mapping using Reactome to fill in gaps and identify events occurring at the cellular/tissue levels. Furthermore, in vivo data from TG-Gates was integrated to finally derive a gene, pathway, biochemical, histopathological and disease network from which specific disease sub-networks can be queried. To test the workflow, non-genotoxic-induced hepatocellular carcinoma (HCC) was selected as a case study. The implementation resulted in the identification of several non-genotoxic-specific HCC-connected genes belonging to cell proliferation, endoplasmic reticulum stress and early apoptosis. Biochemical findings revealed non-genotoxic-specific alkaline phosphatase increase. The explored non-genotoxic-specific histopathology was associated with early stages of hepatic steatosis, transforming into cirrhosis. This work illustrates the utility of computationally predicted constructs in supporting development by using pre-existing knowledge in a fast and unbiased manner.

Integrated analysis of in vitro data and the adverse outcome pathway framework for prioritization and regulatory applications: An exploratory case study using publicly available data on piperonyl butoxide and liver models.

The integration of existing knowledge to support the risk assessment of chemicals is an ongoing challenge for scientists, risk assessors and risk managers. In addition, European Union regulations limiting the use of new animal testing in cosmetics makes already existing information even more valuable. Applying a previous SEURAT-1 program framework to derive predictions of in vivo toxicity responses for a compound, we selected piperonyl butoxide (PBO) as a case study for identification of knowledge and methodology gaps in understanding a compound's effects on the human liver. This is investigated through integration of data from human in vitro transcriptomics studies, biological pathway analysis, chemical and disease associations, and adverse outcome pathway (AOP) information. The outcomes of the analysis are used to generate AOPs of liver-related endpoints, identifying areas of concern for risk assessors and regulators. We demonstrate that integration of data through already existing and publicly available tools can produce outcomes comparable to those that may be found through more conventional time- and resource-intensive methods. It is also expected that, with more refinement, this approach could in the future provide evidence to support chemical risk assessment, while also identifying data gaps for which additional testing may be needed.

Pharmacometabolomic Assessment of Metformin in Non-diabetic, African Americans.

Millions of individuals are diagnosed with type 2 diabetes mellitus (T2D), which increases the risk for a plethora of adverse outcomes including cardiovascular events and kidney disease. Metformin is the most widely prescribed medication for the treatment of T2D; however, its mechanism is not fully understood and individuals vary in their response to this therapy. Here, we use a non-targeted, pharmacometabolomics approach to measure 384 metabolites in 33 non-diabetic, African American subjects dosed with metformin. Three plasma samples were obtained from each subject, one before and two after metformin administration. Validation studies were performed in wildtype mice given metformin. Fifty-four metabolites (including 21 unknowns) were significantly altered upon metformin administration, and 12 metabolites (including six unknowns) were significantly associated with metformin-induced change in glucose (q < 0.2). Of note, indole-3-acetate, a metabolite produced by gut microbes, and 4-hydroxyproline were modulated following metformin exposure in both humans and mice. 2-Hydroxybutanoic acid, a metabolite previously associated with insulin resistance and an early biomarker of T2D, was positively correlated with fasting glucose levels as well as glucose levels following oral glucose tolerance tests after metformin administration. Pathway analysis revealed that metformin administration was associated with changes in a number of metabolites in the urea cycle and in purine metabolic pathways (q < 0.01). Further research is needed to validate the biomarkers of metformin exposure and response identified in this study, and to understand the role of metformin in ammonia detoxification, protein degradation and purine metabolic pathways.

An integrative data mining approach to identifying adverse outcome pathway signatures.

The Adverse Outcome Pathway (AOP) framework is a tool for making biological connections and summarizing key information across different levels of biological organization to connect biological perturbations at the molecular level to adverse outcomes for an individual or population. Computational approaches to explore and determine these connections can accelerate the assembly of AOPs. By leveraging the wealth of publicly available data covering chemical effects on biological systems, computationally-predicted AOPs (cpAOPs) were assembled via data mining of high-throughput screening (HTS) in vitro data, in vivo data and other disease phenotype information. Frequent Itemset Mining (FIM) was used to find associations between the gene targets of ToxCast HTS assays and disease data from Comparative Toxicogenomics Database (CTD) by using the chemicals as the common aggregators between datasets. The method was also used to map gene expression data to disease data from CTD. A cpAOP network was defined by considering genes and diseases as nodes and FIM associations as edges. This network contained 18,283 gene to disease associations for the ToxCast data and 110,253 for CTD gene expression. Two case studies show the value of the cpAOP network by extracting subnetworks focused either on fatty liver disease or the Aryl Hydrocarbon Receptor (AHR). The subnetwork surrounding fatty liver disease included many genes known to play a role in this disease. When querying the cpAOP network with the AHR gene, an interesting subnetwork including glaucoma was identified. While substantial literature exists to support the potential for AHR ligands to elicit glaucoma, it was not explicitly captured in the public annotation information in CTD. The subnetwork from this analysis suggests a cpAOP that includes changes in CYP1B1 expression, which has been previously established in the literature as a primary cause of glaucoma. These case studies highlight the value in integrating multiple data sources when defining cpAOPs for HTS data.

Accelerating Adverse Outcome Pathway Development Using Publicly Available Data Sources.

The adverse outcome pathway (AOP) concept links molecular perturbations with organism and population-level outcomes to support high-throughput toxicity (HTT) testing. International efforts are underway to define AOPs and store the information supporting these AOPs in a central knowledge base; however, this process is currently labor-intensive and time-consuming. Publicly available data sources provide a wealth of information that could be used to define computationally predicted AOPs (cpAOPs), which could serve as a basis for creating expert-derived AOPs in a much more efficient way. Computational tools for mining large datasets provide the means for extracting and organizing the information captured in these public data sources. Using cpAOPs as a starting point for expert-derived AOPs should accelerate AOP development. Coupling this with tools to coordinate and facilitate the expert development efforts will increase the number and quality of AOPs produced, which should play a key role in advancing the adoption of HTT testing, thereby reducing the use of animals in toxicity testing and greatly increasing the number of chemicals that can be tested.

Novel human genetic variants associated with extrapulmonary tuberculosis: a pilot genome wide association study.

BACKGROUND: Approximately 5-10% of persons infected with M. tuberculosis develop tuberculosis, but the factors associated with disease progression are incompletely understood. Both linkage and association studies have identified human genetic variants associated with susceptibility to pulmonary tuberculosis, but few genetic studies have evaluated extrapulmonary disease. Because extrapulmonary and pulmonary tuberculosis likely have different underlying pathophysiology, identification of genetic mutations associated with extrapulmonary disease is important. FINDINGS: We performed a pilot genome-wide association study among 24 persons with previous extrapulmonary tuberculosis and well-characterized immune defects; 24 pulmonary tuberculosis patients and 57 patients with M. tuberculosis infection served as controls. The Affymetrix GeneChip Human Mapping Xba Array was used for genotyping; after careful quality control, genotypes at 44,175 single nucleotide polymorphisms (SNPs) were available for analysis. Eigenstrat quantified population stratification within our sample; logistic regression, using results of the Eigenstrat analysis as a covariate, identified significant associations between groups. Permutation testing controlled the family-wise error rate for each comparison between groups. Four SNPs were significantly associated with extrapulmonary tuberculosis compared to controls with M. tuberculosis infection; one (rs4893980) in the gene PDE11A, one (rs10488286) in KCND2, and one (rs2026414) in PCDH15; one was in chromosome 7 but not associated with a known gene. Two additional variants were significantly associated with extrapulmonary tuberculosis compared with pulmonary tuberculosis; one (rs340708) in the gene FAM135B and one in chromosome 13 but not associated with a known gene. The function of all four genes affects cell signaling and activity, including in the brain. CONCLUSIONS: In this pilot study, we identified 6 novel variants not previously known to be associated with extrapulmonary tuberculosis, including two SNPs more common in persons with extrapulmonary than pulmonary tuberculosis. This provides some support for the hypothesis that the pathogenesis and genetic predisposition to extrapulmonary tuberculosis differs from pulmonary tuberculosis. Further study of these novel SNPs, and more well-powered genome-wide studies of extrapulmonary tuberculosis, is warranted.

Multifactor dimensionality reduction as a filter-based approach for genome wide association studies.

Advances in genotyping technology and the multitude of genetic data available now provide a vast amount of data that is proving to be useful in the quest for a better understanding of human genetic diseases through the study of genetic variation. This has led to the development of approaches such as genome wide association studies (GWAS) designed specifically for interrogating variants across the genome for association with disease, typically by testing single locus, univariate associations. More recently it has been accepted that epistatic (interaction) effects may also be great contributors to these genetic effects, and GWAS methods are now being applied to find epistatic effects. The challenge for these methods still remain in prioritization and interpretation of results, as it has also become standard for initial findings to be independently investigated in replication cohorts or functional studies. This is motivating the development and implementation of filter-based approaches to prioritize variants found to be significant in a discovery stage for follow-up for replication. Such filters must be able to detect both univariate and interactive effects. In the current study we present and evaluate the use of multifactor dimensionality reduction (MDR) as such a filter, with simulated data and a wide range of effect sizes. Additionally, we compare the performance of the MDR filter to a similar filter approach using logistic regression (LR), the more traditional approach used in GWAS analysis, as well as evaporative cooling (EC)-another prominent machine learning filtering method. The results of our simulation study show that MDR is an effective method for such prioritization, and that it can detect main effects, and interactions with or without marginal effects. Importantly, it performed as well as EC and LR for main effect models. It also significantly outperforms LR for various two-locus epistatic models, while it has equivalent results as EC for the epistatic models. The results of this study demonstrate the potential of MDR as a filter to detect gene-gene interactions in GWAS studies.

Polymorphisms in IL-1beta, vitamin D receptor Fok1, and Toll-like receptor 2 are associated with extrapulmonary tuberculosis.

BACKGROUND: Human genetic variants may affect tuberculosis susceptibility, but the immunologic correlates of the genetic variants identified are often unclear. METHODS: We conducted a pilot case-control study to identify genetic variants associated with extrapulmonary tuberculosis in patients with previously characterized immune defects: low CD4+ lymphocytes and low unstimulated cytokine production. Two genetic association approaches were used: 1) variants previously associated with tuberculosis risk; 2) single nucleotide polymorphisms (SNPs) in candidate genes involved in tuberculosis pathogenesis. Single locus association tests and multifactor dimensionality reduction (MDR) assessed main effects and multi-locus interactions. RESULTS: There were 24 extrapulmonary tuberculosis cases (18 black), 24 pulmonary tuberculosis controls (19 black) and 57 PPD+ controls (49 black). In approach 1, 22 SNPs and 3 microsatellites were assessed. In single locus association tests, interleukin (IL)-1beta +3953 C/T was associated with extrapulmonary tuberculosis compared to PPD+ controls (P = 0.049). Among the sub-set of patients who were black, genotype frequencies of the vitamin D receptor (VDR) Fok1 A/G SNP were significantly different in extrapulmonary vs. pulmonary TB patients (P = 0.018). In MDR analysis, the toll-like receptor (TLR) 2 microsatellite had 76% prediction accuracy for extrapulmonary tuberculosis in blacks (P = 0.002). In approach 2, 613 SNPs in 26 genes were assessed. None were associated with extrapulmonary tuberculosis. CONCLUSIONS: In this pilot study among extrapulmonary tuberculosis patients with well-characterized immune defects, genetic variants in IL-1beta, VDR Fok1, and TLR2 were associated with an increased risk of extrapulmonary disease. Additional studies of the underlying mechanism of these genetic variants are warranted.