Implicit Causal Models for Genome-wide Association Studies

By Dustin Tran, David M. Blei. 2017

46 citations recorded by the source corpus.

Progress in probabilistic generative models has accelerated, developing richer models with neural architectures, implicit densities, and with scalable algorithms for their Bayesian inference. However, there has been limited progress in models that capture causal relationships, for example, how individual genetic factors cause major human diseases. In this work, we focus on two challenges in particular: How do we build richer causal models, which can capture highly nonlinear relationships and interactions between multiple causes? How do we adjust for latent confounders, which are variables influencing both cause and effect and which prevent learning of causal relationships? To address these challenges, we synthesize ideas from causality and modern probabilistic modeling. For the first, we describe implicit causal models, a class of causal models that leverages neural architectures with an implicit density. For the second, we describe an implicit causal model that adjusts for confounders by sharing strength across examples. In experiments, we scale Bayesian inference on up to a billion genetic measurements. We achieve state of the art accuracy for identifying causal factors: we significantly outperform existing genetics methods by an absolute difference of 15-45.3%.

Subfields: Artificial Intelligence, Genetics, Molecular Biology

Citations in the corpus

The Blessings of Multiple Causes (2018, unknown): “With a related goal, Tran and Blei (2017) build implicit causal models.”
TensorFlow Distributions (2017, unknown): “…and inference algorithms have enabled new successes in perceptual domains such as images [19], text [4], and audio [48]; and they have advanced scientiﬁc applications such as understanding mouse behavior [23], learning causal factors in genomics [45], and synthesizing new drugs and materials [18].”
Multiple Causal Inference with Latent Confounding (2018, unknown): “Tran and Blei [ 37 ] build implicit models for genome-wide association studies and describe general implicit causal models in the same vein as Kocaoglu et al. [ 19 ] .”
Multiple Causes: A Causal Graphical View (2019, unknown): “Finally, this paper connects to the growing literature on multiple causal inference (Tran & Blei, 2017; Wang & Blei, 2018; Ranganath & Perotte, 2018; Heckerman, 2018; Janzing & Schölkopf, 2018).”
Simple, Distributed, and Accelerated Probabilistic Programming (2018, unknown): “As current work, we are pushing on this design as a stage for fundamental research in generative models and Bayesian neural networks (e.g., [47, 51, 16]).”
Hierarchical Implicit Models and Likelihood-Free Variational Inference (2017, unknown): “Recent work for genomics offers a promising solution [51].”
Multi-cause causal inference with unmeasured confounding and binary outcome. (2019, unknown): “In a recent stream of literature, several authors suggest an alternative approach to this problem by assuming shared confounding between multiple treatments and independence of treatments given the confounder (Wang and Blei, 2019a; Tran and Blei, 2017; Ranganath and Perotte, 2018; Wang and Blei, 201”
Deep Structural Causal Models for Tractable Counterfactual Inference (2020, unknown): “Alternatively, assuming specific causal structures, Tran and Blei [44] and Louizos et al.”