Paper Info
Title
DAGGER: A sequential algorithm for FDR control on DAGs.
Abstract
We propose a top-down algorithm for multiple testing on directed acyclic graphs (DAGs), where nodes represent hypotheses and edges specify a partial ordering in which hypotheses must be tested. The procedure is guaranteed to reject a sub-DAG with bounded false discovery rate (FDR) while satisfying the logical constraint that a rejected nodeu0027s parents must also be rejected. It is designed for sequential testing settings, when the DAG structure is known a priori, but the p-values are obtained selectively (such as sequential conduction of experiments), but the algorithm is also applicable in non-sequential settings when all p-values can be calculated in advance (such as variable/model selection). Our DAGGER algorithm, shorthand for Greedily Evolving Rejections on DAGs, allows for independence, positive or arbitrary dependence of the p-values, and is guaranteed to work on two different types of DAGs: (a) intersection DAGs in which all nodes are intersection hypotheses, with parents being supersets of children, or (b) general DAGs in which all nodes may be elementary hypotheses. The DAGGER procedure has the appealing property that it specializes to known algorithms in the special cases of trees and line graphs, and simplifies to the classic Benjamini-Hochberg procedure when the DAG has no edges. We explore the empirical performance of DAGGER using simulations, as well as a real dataset corresponding to a gene ontology DAG, showing that it performs favorably in terms of time and power.
Year
Venue
Field
2017
arXiv: Methodology
False discovery rate,Line graph,Algorithm,Multiple comparisons problem,Model selection,Directed acyclic graph,Sequential algorithm,Sequential analysis,Mathematics,Bounded function
DocType
Volume
Citations 
Journal
abs/1709.10250
0
PageRank 
References 
Authors
0.34
6
4
Name
Order
Citations
PageRank
Aaditya Ramdas112923.61
Jianbo Chen241.73
Martin J. Wainwright37398533.01
Michael I. Jordan4312203640.80