Using compartmental models to simulate directed acyclic graphs to explore competing causal mechanisms underlying epidemiological study data

Accurately estimating the effect of an exposure on an outcome requires understanding how variables relevant to a study question are causally related to each other. Directed acyclic graphs (DAGs) are used in epidemiology to understand causal processes and determine appropriate statistical approaches to obtain unbiased measures of effect. Compartmental models (CMs) are also used to represent different causal mechanisms, by depicting flows between disease states on the population level. In this paper, we extend a mapping between DAGs and CMs to show how DAG-derived CMs can be used to compare competing causal mechanisms by simulating epidemiological studies and conducting statistical analyses on the simulated data. Through this framework, we can evaluate how robust simulated epidemiological study results are to different biases in study design and underlying causal mechanisms. As a case study, we simulated a longitudinal cohort study to examine the obesity paradox: the apparent protective effect of obesity on mortality among diabetic ever-smokers, but not among diabetic never-smokers. Our simulations illustrate how study design bias (e.g. reverse causation), can lead to the obesity paradox. Ultimately, we show the utility of transforming DAGs into in silico laboratories within which researchers can systematically evaluate bias, and inform analyses and study design.

Download Full-text

Mathematical modeling of directed acyclic graphs to explore competing causal mechanisms underlying epidemiological study data

10.1101/19007922 ◽

2019 ◽

Author(s):

Joshua Havumaki ◽

Marisa C. Eisenberg

Keyword(s):

Epidemiological Study ◽

Study Design ◽

Simulated Data ◽

Population Level ◽

Obesity Paradox ◽

Epidemiological Studies ◽

Directed Acyclic Graphs ◽

Causal Mechanisms ◽

Acyclic Graphs ◽

Study Results

1AbstractAccurately estimating the effect of an exposure on an outcome requires understanding how variables relevant to a study question are causally related to each other. Directed acyclic graphs (DAGs) are used in epidemiology to understand causal processes and determine appropriate statistical approaches to obtain unbiased measures of effect. Compartmental models (CMs) are also used to represent different causal mechanisms, by depicting flows between disease states on the population level. In this paper, we extend a mapping between DAGs and CMs to show how DAG–derived CMs can be used to compare competing causal mechanisms by simulating epidemiological studies and conducting statistical analyses on the simulated data. Through this framework, we can evaluate how robust simulated epidemiological study results are to different biases in study design and underlying causal mechanisms. As a case study, we simulated a longitudinal cohort study to examine the obesity paradox: the apparent protective effect of obesity on mortality among diabetic ever-smokers, but not among diabetic never-smokers. Our simulations illustrate how study design bias (e.g., reverse causation), can lead to the obesity paradox. Ultimately, we show the utility of transforming DAGs into in silico laboratories within which researchers can systematically evaluate bias, and inform analyses and study design.

Download Full-text

ANALYZING MICROARRAY DATA WITH TRANSITIVE DIRECTED ACYCLIC GRAPHS

Journal of Bioinformatics and Computational Biology ◽

10.1142/s0219720009003972 ◽

2009 ◽

Vol 07 (01) ◽

pp. 135-156 ◽

Cited By ~ 7

Author(s):

VINHTHUY PHAN ◽

E. OLUSEGUN GEORGE ◽

QUYNH T. TRAN ◽

SHIRLEAN GOODWIN ◽

SRIDEVI BODREDDIGARI ◽

...

Keyword(s):

Treatment Effects ◽

Simulated Data ◽

Directed Acyclic Graphs ◽

Controlled Study ◽

Clustering Methods ◽

Microarray Experiments ◽

Structure Activity ◽

Acyclic Graphs ◽

Multiple Treatments ◽

Post Hoc

Post hoc assignment of patterns determined by all pairwise comparisons in microarray experiments with multiple treatments has been proven to be useful in assessing treatment effects. We propose the usage of transitive directed acyclic graphs (tDAG) as the representation of these patterns and show that such representation can be useful in clustering treatment effects, annotating existing clustering methods, and analyzing sample sizes. Advantages of this approach include: (1) unique and descriptive meaning of each cluster in terms of how genes respond to all pairs of treatments; (2) insensitivity of the observed patterns to the number of genes analyzed; and (3) a combinatorial perspective to address the sample size problem by observing the rate of contractible tDAG as the number of replicates increases. The advantages and overall utility of the method in elaborating drug structure activity relationships are exemplified in a controlled study with real and simulated data.

Download Full-text

Directed acyclic graphs to explore causality in Epidemiological study designs, part I: an introduction to DAGs

Qeios ◽

10.32388/ffh3gu ◽

2020 ◽

Author(s):

Arindam Basu

Keyword(s):

Epidemiological Study ◽

Directed Acyclic Graphs ◽

Acyclic Graphs ◽

Study Designs

Download Full-text

Cancer progression models and fitness landscapes: a many-to-many relationship

10.1101/141465 ◽

2017 ◽

Author(s):

Ramon Diaz-Uriarte

Keyword(s):

Cancer Progression ◽

Fitness Landscape ◽

Simulated Data ◽

Directed Acyclic Graphs ◽

Fitness Landscapes ◽

Gene Interactions ◽

Cross Sectional ◽

Data Set ◽

Large Variability ◽

Acyclic Graphs

AbstractThe identification of constraints, due to gene interactions, in the order of accumulation of mutations during cancer progression can allow us to single out therapeutic targets. Cancer progression models (CPMs) use genotype frequency data from cross-sectional samples to try to identify these constraints, and return Directed Acyclic Graphs (DAGs) of genes. On the other hand, fitness landscapes, which map genotypes to fitness, contain all possible paths of tumor progression. Thus, we expect a correspondence between DAGs from CPMs and the fitness landscapes where evolution happened. But many fitness landscapes —e.g., those with reciprocal sign epistasis— cannot be represented by CPMs. Using simulated data under 500 fitness landscapes, I show that CPMs’ performance (prediction of genotypes that can exist) degrades with reciprocal sign epistasis. There is large variability in the DAGs inferred from each landscape, which is also affected by mutation rate, detection regime, and fitness landscape features, in ways that depend on CPM method. And the same DAG is often observed in very different landscapes, which differ in more than 50% of their accessible genotypes. Using a pancreatic data set, I show that this many-to-many relationship affects the analysis of empirical data. Fitness landscapes that are widely different from each other can, when evolutionary processes run repeatedly on them, both produce data similar to the empirically observed one, and lead to DAGs that are very different among themselves. Because reciprocal sign epistasis can be common in cancer, these results question the use and interpretation of CPMs.

Download Full-text

Adolescent Paranoia: Prevalence, Structure, and Causal Mechanisms

Schizophrenia Bulletin ◽

10.1093/schbul/sby180 ◽

2018 ◽

Vol 45 (5) ◽

pp. 1134-1142 ◽

Cited By ~ 5

Author(s):

Jessica C Bird ◽

Robin Evans ◽

Felicity Waite ◽

Bao S Loe ◽

Daniel Freeman

Keyword(s):

Social Media ◽

Negative Affect ◽

Peer Relationships ◽

Peer Acceptance ◽

Continuous Distribution ◽

Directed Acyclic Graphs ◽

Causal Mechanisms ◽

Acyclic Graphs ◽

Physical Threat ◽

Fear Of Rejection

AbstractBackgroundAdolescence can be a challenging time, characterized by self-consciousness, heightened regard for peer acceptance, and fear of rejection. Interpersonal concerns are amplified by unpredictable social interactions, both online and offline. This developmental and social context is potentially conducive to the emergence of paranoia. However, research on paranoia during adolescence is scarce.MethodOur aim was to examine the prevalence, structure, and probabilistic causal mechanisms of adolescent paranoia. A representative school cohort of 801 adolescents (11–15 y) completed measures of paranoia and a range of affective, cognitive, and social factors. A Bayesian approach with Directed Acyclic Graphs (DAGs) was used to assess the causal interactions with paranoia.ResultsParanoid thoughts were very common, followed a continuous distribution, and were hierarchically structured. There was an overall paranoia factor, with sub-factors of social fears, physical threat fears, and conspiracy concerns. With all other variables controlled, DAG analysis identified paranoia had dependent relationships with negative affect, peer difficulties, bullying, and cognitive-affective responses to social media. The causal directions could not be fully determined, but it was more likely that negative affect contributed to paranoia and paranoia impacted peer relationships. Problematic social media use did not causally influence paranoia.ConclusionsThere is a continuum of paranoia in adolescence and occasional suspicions are common at this age. Anxiety and depression are closely connected with paranoia and may causally contribute to its development. Paranoia may negatively impact adolescent peer relationships. The clinical significance of paranoia in adolescents accessing mental health services must now be established.

Download Full-text