DiscoBAX: Discovery of optimal intervention sets in genomic experiment design

Clare Lyle, Arash Mehrjou, Pascal Notin, Andrew Jesson, Stefan Bauer, Yarin Gal, Patrick Schwab
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:23170-23189, 2023.

Abstract

The discovery of therapeutics to treat genetically-driven pathologies relies on identifying genes involved in the underlying disease mechanism. Existing approaches search over the billions of potential interventions to maximize the expected influence on the target phenotype. However, to reduce the risk of failure in future stages of trials, practical experiment design aims to find a set of interventions that maximally change a target phenotype via diverse mechanisms. We propose DiscoBAX - a sample-efficient method for maximizing the rate of significant discoveries per experiment while simultaneously probing for a wide range of diverse mechanisms during a genomic experiment campaign. We provide theoretical guarantees of optimality under standard assumptions, and conduct a comprehensive experimental evaluation covering both synthetic as well as real-world experimental design tasks. DiscoBAX outperforms existing state-of-the-art methods for experimental design, selecting effective and diverse perturbations in biological systems.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-lyle23a, title = {{D}isco{BAX}: Discovery of optimal intervention sets in genomic experiment design}, author = {Lyle, Clare and Mehrjou, Arash and Notin, Pascal and Jesson, Andrew and Bauer, Stefan and Gal, Yarin and Schwab, Patrick}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {23170--23189}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/lyle23a/lyle23a.pdf}, url = {https://proceedings.mlr.press/v202/lyle23a.html}, abstract = {The discovery of therapeutics to treat genetically-driven pathologies relies on identifying genes involved in the underlying disease mechanism. Existing approaches search over the billions of potential interventions to maximize the expected influence on the target phenotype. However, to reduce the risk of failure in future stages of trials, practical experiment design aims to find a set of interventions that maximally change a target phenotype via diverse mechanisms. We propose DiscoBAX - a sample-efficient method for maximizing the rate of significant discoveries per experiment while simultaneously probing for a wide range of diverse mechanisms during a genomic experiment campaign. We provide theoretical guarantees of optimality under standard assumptions, and conduct a comprehensive experimental evaluation covering both synthetic as well as real-world experimental design tasks. DiscoBAX outperforms existing state-of-the-art methods for experimental design, selecting effective and diverse perturbations in biological systems.} }
Endnote
%0 Conference Paper %T DiscoBAX: Discovery of optimal intervention sets in genomic experiment design %A Clare Lyle %A Arash Mehrjou %A Pascal Notin %A Andrew Jesson %A Stefan Bauer %A Yarin Gal %A Patrick Schwab %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-lyle23a %I PMLR %P 23170--23189 %U https://proceedings.mlr.press/v202/lyle23a.html %V 202 %X The discovery of therapeutics to treat genetically-driven pathologies relies on identifying genes involved in the underlying disease mechanism. Existing approaches search over the billions of potential interventions to maximize the expected influence on the target phenotype. However, to reduce the risk of failure in future stages of trials, practical experiment design aims to find a set of interventions that maximally change a target phenotype via diverse mechanisms. We propose DiscoBAX - a sample-efficient method for maximizing the rate of significant discoveries per experiment while simultaneously probing for a wide range of diverse mechanisms during a genomic experiment campaign. We provide theoretical guarantees of optimality under standard assumptions, and conduct a comprehensive experimental evaluation covering both synthetic as well as real-world experimental design tasks. DiscoBAX outperforms existing state-of-the-art methods for experimental design, selecting effective and diverse perturbations in biological systems.
APA
Lyle, C., Mehrjou, A., Notin, P., Jesson, A., Bauer, S., Gal, Y. & Schwab, P.. (2023). DiscoBAX: Discovery of optimal intervention sets in genomic experiment design. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:23170-23189 Available from https://proceedings.mlr.press/v202/lyle23a.html.

Related Material