Deep neural networks with knockoff features identify nonlinear causal relations and estimate effect sizes in complex biological systems
GigaScience · 2022
DOI: 10.1093/gigascience/giad044 ↗Kurz gefasst
Fachartikel in der Zeitschrift GigaScience, 2022. Verfasst von Zhenjiang Fan und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, University of Pittsburgh, Pittsburgh, PA 15213, USA. Hauptthemen laut MEDLINE: Neural Networks, Computer. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 122 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NHLBI NIH HHS, NIGMS NIH HHS, NCI NIH HHS.
Abstract (Originalsprache)
Background
Learning the causal structure helps identify risk factors, disease mechanisms, and candidate therapeutics for complex diseases. However, although complex biological systems are characterized by nonlinear associations, existing bioinformatic methods of causal inference cannot identify the nonlinear relationships and estimate their effect size.
Results
To overcome these limitations, we developed the first computational method that explicitly learns nonlinear causal relations and estimates the effect size using a deep neural network approach coupled with the knockoff framework, named causal directed acyclic graphs using deep learning variable selection (DAG-deepVASE). Using simulation data of diverse scenarios and identifying known and novel causal relations in molecular and clinical data of various diseases, we demonstrated that DAG-deepVASE consistently outperforms existing methods in identifying true and known causal relations. In the analyses, we also illustrate how identifying nonlinear causal relations and estimating their effect size help understand the complex disease pathobiology, which is not possible using other methods.
Conclusions
With these advantages, the application of DAG-deepVASE can help identify driver genes and therapeutic agents in biomedical studies and clinical trials.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Computer Science, University of Pittsburgh, Pittsburgh, PA 15213, USA.
- Division of Pediatric Critical Care Medicine, Department of Critical Care Medicine, Children's Hospital of Pittsburgh, Center for Critical Care Nephrology and Clinical Research Investigation and Systems Modeling of Acut…
- Department of Human Genetics, University of Pittsburgh, Pittsburgh, PA 15213, USA.
- Department of Epidemiology, University of Florida, Gainesville, FL 32610, USA.
- Pediatric Rheumatology, The Children's Hospital of Philadelphia, Philadelphia, PA 19104, USA.
- Division of Pediatric Pulmonary Medicine, Children's Hospital of Pittsburgh, Pittsburgh, PA 15224, USA.
ORCID-Profile: Zhenjiang Fan, Kate F Kernan, Aditya Sriram, Panayiotis V Benos, Scott W Canna, Soyeon Kim, Hyun Jung Park
Förderung
- NHLBI NIH HHS
- NIGMS NIH HHS
- NCI NIH HHS
- NICHD NIH HHS
- University of Pittsburgh
- National Institutes of Health
- Eunice Kennedy Shriver National Institutes of Child Health and Human Development
- National Heart, Lung, and Blood Institute
Zitieren
APA
Fan, Z., Kernan, K. F., Sriram, A., Benos, P. V., Canna, S. W., Carcillo, J. A., Kim, S., & Park, H. J. (2022). Deep neural networks with knockoff features identify nonlinear causal relations and estimate effect sizes in complex biological systems. GigaScience. https://doi.org/10.1093/gigascience/giad044
Vancouver
Fan Z, Kernan KF, Sriram A, Benos PV, Canna SW, Carcillo JA, et al. Deep neural networks with knockoff features identify nonlinear causal relations and estimate effect sizes in complex biological systems. GigaScience. 2022. doi:10.1093/gigascience/giad044
BibTeX
@article{Fan2022,
author = {Fan, Zhenjiang and Kernan, Kate F and Sriram, Aditya and Benos, Panayiotis V and Canna, Scott W and Carcillo, Joseph A and Kim, Soyeon and Park, Hyun Jung},
title = {Deep neural networks with knockoff features identify nonlinear causal relations and estimate effect sizes in complex biological systems},
journal = {GigaScience},
year = {2022},
doi = {10.1093/gigascience/giad044}
}
EUVIMED