Abstract
Learning the causal relationships that define a molecular system allows us to predict how the system will respond to different interventions. Distinguishing causality from mere association typically requires randomized experiments. Methods for automated causal discovery from limited experiments exist, but have so far rarely been tested in systems biology applications. In this work, we apply state-of-the art causal discovery methods on a large collection of public mass cytometry data sets, measuring intra-cellular signaling proteins of the human immune system and their response to several perturbations. We show how different experimental conditions can be used to facilitate causal discovery, and apply two fundamental methods that produce context-specific causal predictions. Causal predictions were reproducible across independent data sets from two different studies, but often disagree with the KEGG pathway databases. Within this context, we discuss the caveats we need to overcome for automated causal discovery to become a part of the routine data analysis in systems biology.
Highlights
Computational causality has developed a language to describe, quantify and reason with causal claims
As a test bed dataset to assess the applicability and performance of automated causal discovery methods, we used a public collection of mass cytometry data[5]
Multiplexed mass cytometry was used to measure the abundance of surface proteins and intracellular phosphorylated proteins in human peripheral blood mononuclear cells (PBMCs) that were stimulated with different factors in vitro
Summary
Computational causality has developed a language to describe, quantify and reason with causal claims. In a seminal paper for applied causal discovery[4], the authors were able to almost flawlessly reconstruct a known causal signaling pathway from a mixture of experimental and observational flow cytometry measurements. Mass cytometry is a technique that can be used to singularize cells and measure protein abundance on the cellular level, resulting in very large sample sizes that are suitable for causal discovery methods. Code for reproducing the results is available in https://github.com/mensxmachina/Mass-Cytometry These results indicate that (a) de novo discovery of causal pathway relations is still a challenging task for current causal discovery methods, despite the previous positive results[4] and (b) current causal discovery methods do identify reproducible findings across similar data sets. There results constitute an important step for further developments in order to apply causal discovery methods successfully to biological single cell data in the future
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
