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

• DOI: 10.1038/s41587-022-01520-x
• Publication Date: 2023-01-02T17:02:56Z
• Authors: Rosa Lundbye Allesøe; Agnete Troen Lund...; Ricardo Hernández...; Alejandro Aguayo-...; Joachim Johansen; Jakob Nybo Nissen; Caroline Brorsson; Gianluca Mazzoni; Lili Niu; Jorge Hernansanz ...; Cristina Leal Rod...; Valentas Brasas; Henry Webel; Michael Eriksen B...; Anders Gorm Pedersen; Piotr Jaroslaw Ch...; Ulrik Plesner Jac...; Andrea Mari; Robert Koivula; Anubha Mahajan; Ana Vinuela; Juan Fernandez Tajes; Sapna Sharma; Mark Haid; Mun-Gwan Hong; Petra B. Musholt; Federico De Masi; Josef Vogt; Helle Krogh Pedersen; Valborg Gudmundsd...; Angus Jones; Gwen Kennedy; Jimmy Bell; E. Louise Thomas; Gary Frost; Henrik Thomsen; Elizaveta Hansen; Tue Haldor Hansen; Henrik Vestergaard; Mirthe Muilwijk; Marieke T. Blom; Leen M. ‘t Hart; Francois Pattou; Violeta Raverdy; Soren Brage; Tarja Kokkola; Alison Heggie; Donna McEvoy; Miranda Mourby; Jane Kaye; Andrew Hattersley; Timothy McDonald; Martin Ridderstråle; Mark Walker; Ian Forgie; Giuseppe N. Giordano; Imre Pavo; Hartmut Ruetten; Oluf Pedersen; Torben Hansen; Emmanouil Dermitz...; Paul W. Franks; Jochen M. Schwenk; Jerzy Adamski; Mark I. McCarthy; Ewan Pearson; Karina Banasik; Simon Rasmussen; Søren Brunak; ; Philippe Froguel; Cecilia Engel Thomas; Ragna Haussler; Joline Beulens; Femke Rutters; Giel Nijpels; Sabine van Oort; Lenka Groeneveld; Petra Elders; Toni Giorgino; Marianne Rodriquez; Rachel Nice; Mandy Perry; Susanna Bianzano; Ulrike Graefe-Mody; Anita Hennige; Rolf Grempler; Patrick Baum; Hans-Henrik Stærf...; Nisha Shah; Harriet Teare; Beate Ehrhardt; Joachim Tillner; Christiane Dings; Thorsten Lehr; Nina Scherer; Iryna Sihinevich; Louise Cabrelli; Heather Loftus; Roberto Bizzotto; Andrea Tura; Koen Dekkers; Nienke van Leeuwen; Leif Groop; Roderick Slieker; Anna Ramisch; Christopher Jennison; Ian McVittie; Francesca Frau; Birgit Steckel-Ha...; Kofi Adragni; Melissa Thomas; Naeimeh Atabaki P...; Hugo Fitipaldi; Azra Kurbasic; Pascal Mutie; Hugo Pomares-Millan; Amelie Bonnefond; Mickael Canouil; Robert Caiazzo; Helene Verkindt; Reinhard Holl; Teemu Kuulasmaa; Harshal Deshmukh; Henna Cederberg; Markku Laakso; Jagadish Vangipurapu; Matilda Dale; Barbara Thorand; Claudia Nicolay; Andreas Fritsche; Anita Hill; Michelle Hudson; Claire Thorne; Kristine Allin; Manimozhiyan Arum...; Anna Jonsson; Line Engelbrechtsen; Annemette Forman; Avirup Dutta; Nadja Sondertoft; Yong Fan; Stephen Gough; Neil Robertson; Nicky McRobert; Agata Wesolowska-...; Andrew Brown; David Davtian; Adem Dawed; Louise Donnelly; Colin Palmer; Margaret White; Jorge Ferrer; Brandon Whitcher; Anna Artati; Cornelia Prehn; Jonathan Adam; Harald Grallert; Ramneek Gupta; Peter Wad Sackett; Birgitte Nilsson; Konstantinos Tsir...; Rebeca Eriksen; Bernd Jablonka; Mathias Uhlen; Johann Gassenhuber; Tania Baltauss; Nathalie de Preville; Maria Klintenberg; Moustafa Abdalla
• Keywords: /dk/atira/pure/subjectarea/asjc/1300/1313; /dk/atira/pure/subjectarea/asjc/2200/2204; [SDV]Life Sciences [q-bio]; 610; Article; name=Applied Microbiology and Biotechnology; 03 medical and health sciences; Deep Learning; Machine learning; name=Molecular Medicine; Humans; /dk/atira/pure/subjectarea/asjc/2400/2402; /dk/atira/pure/sustainabledevelopmentgoals/good_health_and_well_being; name=SDG 3 - Good Health and Well-being; IMI DIRECT Consortium; name=Biomedical Engineering; generative deep-learning models; 0303 health sciences; name=Bioengineering; name=Biotechnology; Type 2 diabetes; 3. Good health; Diabetes Mellitus, Type 2; Data integration; Systems biology; /dk/atira/pure/subjectarea/asjc/1500/1502; Algorithms; /dk/atira/pure/subjectarea/asjc/1300/1305

AbstractThe 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.