Titre
Multimodal learning in clinical proteomics: enhancing antimicrobial resistance prediction models with chemical information.
Type
article
Institution
UNIL/CHUV/Unisanté + institutions partenaires
Périodique
Auteur(s)
Visonà, G.
Auteure/Auteur
Duroux, D.
Auteure/Auteur
Miranda, L.
Auteure/Auteur
Sükei, E.
Auteure/Auteur
Li, Y.
Auteure/Auteur
Borgwardt, K.
Auteure/Auteur
Oliver, C.
Auteure/Auteur
Liens vers les unités
ISSN
1367-4811
Statut éditorial
Publié
Date de publication
2023-12-01
Volume
39
Numéro
12
Première page
btad717
Peer-reviewed
Oui
Langue
anglais
Notes
Publication types: Journal Article ; Research Support, Non-U.S. Gov't
Publication Status: ppublish
Publication Status: ppublish
Résumé
Large-scale clinical proteomics datasets of infectious pathogens, combined with antimicrobial resistance outcomes, have recently opened the door for machine learning models which aim to improve clinical treatment by predicting resistance early. However, existing prediction frameworks typically train a separate model for each antimicrobial and species in order to predict a pathogen's resistance outcome, resulting in missed opportunities for chemical knowledge transfer and generalizability.
We demonstrate the effectiveness of multimodal learning over proteomic and chemical features by exploring two clinically relevant tasks for our proposed deep learning models: drug recommendation and generalized resistance prediction. By adopting this multi-view representation of the pathogenic samples and leveraging the scale of the available datasets, our models outperformed the previous single-drug and single-species predictive models by statistically significant margins. We extensively validated the multi-drug setting, highlighting the challenges in generalizing beyond the training data distribution, and quantitatively demonstrate how suitable representations of antimicrobial drugs constitute a crucial tool in the development of clinically relevant predictive models.
The code used to produce the results presented in this article is available at https://github.com/BorgwardtLab/MultimodalAMR.
We demonstrate the effectiveness of multimodal learning over proteomic and chemical features by exploring two clinically relevant tasks for our proposed deep learning models: drug recommendation and generalized resistance prediction. By adopting this multi-view representation of the pathogenic samples and leveraging the scale of the available datasets, our models outperformed the previous single-drug and single-species predictive models by statistically significant margins. We extensively validated the multi-drug setting, highlighting the challenges in generalizing beyond the training data distribution, and quantitatively demonstrate how suitable representations of antimicrobial drugs constitute a crucial tool in the development of clinically relevant predictive models.
The code used to produce the results presented in this article is available at https://github.com/BorgwardtLab/MultimodalAMR.
PID Serval
serval:BIB_A4D8C2BA1377
PMID
Open Access
Oui
Date de création
2024-01-12T07:59:57.383Z
Date de création dans IRIS
2025-05-20T23:39:49Z
Fichier(s)![Vignette d'image]()
En cours de chargement...
Nom
38001023_BIB_A4D8C2BA1377.pdf
Version du manuscrit
published
Licence
https://creativecommons.org/licenses/by/4.0
Taille
2.41 MB
Format
Adobe PDF
PID Serval
serval:BIB_A4D8C2BA1377.P001
URN
urn:nbn:ch:serval-BIB_A4D8C2BA13770
Somme de contrôle
(MD5):ec4a9ebbb7e417d07ea38f857c065861