Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Discovery of drug–omics associations in type 2 diabetes with generative deep-learning models

Allesøe, R.L. ; Ridderstråle, M. LU ; Giordano, G.N. LU ; Franks, P.W. LU ; Groop, L. LU ; Pasdar, N.A. ; Fitipaldi, H. LU ; Kurbasic, A. LU ; Mutie, P. LU and Pomares-Millan, H. LU orcid , et al. (2023) In Nature Biotechnology
Abstract
The application of multiple omics technologies in biomedical cohorts has the potential to reveal patient-level disease characteristics and individualized response to treatment. However, the scale and heterogeneous nature of multi-modal data makes integration and inference a non-trivial task. We developed a deep-learning-based framework, multi-omics variational autoencoders (MOVE), to integrate such data and applied it to a cohort of 789 people with newly diagnosed type 2 diabetes with deep multi-omics phenotyping from the DIRECT consortium. Using in silico perturbations, we identified drug–omics associations across the multi-modal datasets for the 20 most prevalent drugs given to people with type 2 diabetes with substantially higher... (More)
The application of multiple omics technologies in biomedical cohorts has the potential to reveal patient-level disease characteristics and individualized response to treatment. However, the scale and heterogeneous nature of multi-modal data makes integration and inference a non-trivial task. We developed a deep-learning-based framework, multi-omics variational autoencoders (MOVE), to integrate such data and applied it to a cohort of 789 people with newly diagnosed type 2 diabetes with deep multi-omics phenotyping from the DIRECT consortium. Using in silico perturbations, we identified drug–omics associations across the multi-modal datasets for the 20 most prevalent drugs given to people with type 2 diabetes with substantially higher sensitivity than univariate statistical tests. From these, we among others, identified novel associations between metformin and the gut microbiota as well as opposite molecular responses for the two statins, simvastatin and atorvastatin. We used the associations to quantify drug–drug similarities, assess the degree of polypharmacy and conclude that drug effects are distributed across the multi-omics modalities. © 2023, The Author(s). (Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; ; and , et al. (More)
; ; ; ; ; ; ; ; ; ; and (Less)
author collaboration
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Deep learning, Learning systems, Patient treatment, 'omics', Auto encoders, In-silico, Learning models, Multi-modal data, Multi-modal dataset, Non-trivial tasks, Omics technologies, Phenotyping, Type-2 diabetes, Modal analysis
in
Nature Biotechnology
publisher
Nature Publishing Group
external identifiers
  • scopus:85145508974
  • pmid:36593394
ISSN
1087-0156
DOI
10.1038/s41587-022-01520-x
language
English
LU publication?
yes
id
13c6b992-22aa-4bcc-b353-dbd6f2f72642
date added to LUP
2023-01-16 11:19:14
date last changed
2024-05-02 22:07:16
@article{13c6b992-22aa-4bcc-b353-dbd6f2f72642,
  abstract     = {{The application of multiple omics technologies in biomedical cohorts has the potential to reveal patient-level disease characteristics and individualized response to treatment. However, the scale and heterogeneous nature of multi-modal data makes integration and inference a non-trivial task. We developed a deep-learning-based framework, multi-omics variational autoencoders (MOVE), to integrate such data and applied it to a cohort of 789 people with newly diagnosed type 2 diabetes with deep multi-omics phenotyping from the DIRECT consortium. Using in silico perturbations, we identified drug–omics associations across the multi-modal datasets for the 20 most prevalent drugs given to people with type 2 diabetes with substantially higher sensitivity than univariate statistical tests. From these, we among others, identified novel associations between metformin and the gut microbiota as well as opposite molecular responses for the two statins, simvastatin and atorvastatin. We used the associations to quantify drug–drug similarities, assess the degree of polypharmacy and conclude that drug effects are distributed across the multi-omics modalities. © 2023, The Author(s).}},
  author       = {{Allesøe, R.L. and Ridderstråle, M. and Giordano, G.N. and Franks, P.W. and Groop, L. and Pasdar, N.A. and Fitipaldi, H. and Kurbasic, A. and Mutie, P. and Pomares-Millan, H. and Klintenberg, M. and Abdalla, M.}},
  issn         = {{1087-0156}},
  keywords     = {{Deep learning; Learning systems; Patient treatment; 'omics'; Auto encoders; In-silico; Learning models; Multi-modal data; Multi-modal dataset; Non-trivial tasks; Omics technologies; Phenotyping; Type-2 diabetes; Modal analysis}},
  language     = {{eng}},
  publisher    = {{Nature Publishing Group}},
  series       = {{Nature Biotechnology}},
  title        = {{Discovery of drug–omics associations in type 2 diabetes with generative deep-learning models}},
  url          = {{http://dx.doi.org/10.1038/s41587-022-01520-x}},
  doi          = {{10.1038/s41587-022-01520-x}},
  year         = {{2023}},
}