Drug discovery under covariate shift with domain-informed prior distributions over functions

Accelerating the discovery of novel and more effective therapeutics is an important pharmaceutical problem in which deep learning is playing an increasingly significant role. However, real-world drug discovery tasks are often characterized by a scarcity of labeled data and significant covariate shif...

Full description

Bibliographic Details
Main Authors: Klarner, L, Rudner, T, Reutlinger, M, Schindler, T, Morris, G, Deane, CM, Yeh, YW
Other Authors: Krause, A
Format: Conference item
Language:English
Published: Journal of Machine Learning Research 2023
_version_ 1797110739450724352
author Klarner, L
Rudner, T
Reutlinger, M
Schindler, T
Morris, G
Deane, CM
Yeh, YW
author2 Krause, A
author_facet Krause, A
Klarner, L
Rudner, T
Reutlinger, M
Schindler, T
Morris, G
Deane, CM
Yeh, YW
author_sort Klarner, L
collection OXFORD
description Accelerating the discovery of novel and more effective therapeutics is an important pharmaceutical problem in which deep learning is playing an increasingly significant role. However, real-world drug discovery tasks are often characterized by a scarcity of labeled data and significant covariate shift—a setting that poses a challenge to standard deep learning methods. In this paper, we present Q-SAVI, a probabilistic model able to address these challenges by encoding explicit prior knowledge of the data-generating process into a prior distribution over functions, presenting researchers with a transparent and probabilistically principled way to encode data-driven modeling preferences. Building on a novel, gold-standard bioactivity dataset that facilitates a meaningful comparison of models in an extrapolative regime, we explore different approaches to induce data shift and construct a challenging evaluation setup. We then demonstrate that using Q-SAVI to integrate contextualized prior knowledge of drug-like chemical space into the modeling process affords substantial gains in predictive accuracy and calibration, outperforming a broad range of state-of-the-art self-supervised pre-training and domain adaptation techniques.
first_indexed 2024-03-07T08:00:31Z
format Conference item
id oxford-uuid:733eaa50-f835-4ead-adc4-9e76a4c79c46
institution University of Oxford
language English
last_indexed 2024-03-07T08:00:31Z
publishDate 2023
publisher Journal of Machine Learning Research
record_format dspace
spelling oxford-uuid:733eaa50-f835-4ead-adc4-9e76a4c79c462023-09-19T08:20:44ZDrug discovery under covariate shift with domain-informed prior distributions over functionsConference itemhttp://purl.org/coar/resource_type/c_5794uuid:733eaa50-f835-4ead-adc4-9e76a4c79c46EnglishSymplectic ElementsJournal of Machine Learning Research2023Klarner, LRudner, TReutlinger, MSchindler, TMorris, GDeane, CMYeh, YWKrause, ABrunskill, ECho, KEngelhardt, BSabato, SScarlett, JAccelerating the discovery of novel and more effective therapeutics is an important pharmaceutical problem in which deep learning is playing an increasingly significant role. However, real-world drug discovery tasks are often characterized by a scarcity of labeled data and significant covariate shift—a setting that poses a challenge to standard deep learning methods. In this paper, we present Q-SAVI, a probabilistic model able to address these challenges by encoding explicit prior knowledge of the data-generating process into a prior distribution over functions, presenting researchers with a transparent and probabilistically principled way to encode data-driven modeling preferences. Building on a novel, gold-standard bioactivity dataset that facilitates a meaningful comparison of models in an extrapolative regime, we explore different approaches to induce data shift and construct a challenging evaluation setup. We then demonstrate that using Q-SAVI to integrate contextualized prior knowledge of drug-like chemical space into the modeling process affords substantial gains in predictive accuracy and calibration, outperforming a broad range of state-of-the-art self-supervised pre-training and domain adaptation techniques.
spellingShingle Klarner, L
Rudner, T
Reutlinger, M
Schindler, T
Morris, G
Deane, CM
Yeh, YW
Drug discovery under covariate shift with domain-informed prior distributions over functions
title Drug discovery under covariate shift with domain-informed prior distributions over functions
title_full Drug discovery under covariate shift with domain-informed prior distributions over functions
title_fullStr Drug discovery under covariate shift with domain-informed prior distributions over functions
title_full_unstemmed Drug discovery under covariate shift with domain-informed prior distributions over functions
title_short Drug discovery under covariate shift with domain-informed prior distributions over functions
title_sort drug discovery under covariate shift with domain informed prior distributions over functions
work_keys_str_mv AT klarnerl drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions
AT rudnert drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions
AT reutlingerm drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions
AT schindlert drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions
AT morrisg drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions
AT deanecm drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions
AT yehyw drugdiscoveryundercovariateshiftwithdomaininformedpriordistributionsoverfunctions