Available via license: CC BY 4.0
Content may be subject to copyright.
Nature Biotechnology
nature biotechnology
https://doi.org/10.1038/s41587-022-01520-xArticle
Discovery of drug–omics associations
in type 2 diabetes with generative
deep-learning models
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 identied 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, identied 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 eects are distributed across the
multi-omics modalities.
Drug-response patterns in individuals with complex disease, such as
type 2 diabetes (T2D), are intricate. Multiple organs and confounders
are typically involved including comorbidities and polypharmacy
1,2
.
Conversely, treatment with one or more drugs and the associated poly-
pharmacy effects can have considerable impact on the molecular profile
of the individual; however, such changes are still largely unknown
3
. The
increasing availability of deep phenotyping and multi-omics screening
has proven to be beneficial in the characterization of T2D and other
diseases4–7, and offer the opportunity to gain mechanistic insights on
the action of drugs on disease processes.
Cohort studies can be highly useful for investigating associa-
tions between drugs and molecular phenotypes, and can be used to
tailor the design of randomized control studies to assess direct causal
relationships8. Common approaches to analysis of cohort data apply
univariate statistical methods, linear and logistic regression, dimen-
sionality reduction and clustering analyses. However, when expanding
to multi-omics data such analyses are not straightforward and tradi-
tional methods of data interpretation are insufficient to exploit the full
scope of multi-modality data.
Here we investigate vertical data integration, where multiple omics
datasets have been generated for the same samples. Challenges that
must be overcome include integration of data across multiple con-
tinuous and discrete data modalities, efficient handling of missing
data or even large missing parts of specific data types, differences in
dimensionality, modality-specific noise and how to extract associations
across data modalities
9–11
. There are several strategies for vertical inte-
gration of multi-modal datasets, such as element-wise addition of one
dataset at a time, learning individual representations for each dataset
before fusion, or multi-dimensional fusion where representations are
learned from the input data altogether
9,12–14
. Examples are multi-omics
factor analysis (MOFA), iCluster, and data integration analysis for
biomarker discovery using latent components (DIABLO) implemented
Received: 15 March 2022
Accepted: 20 September 2022
Published online: xx xx xxxx
Check for updates
e-mail: simon.rasmussen@cpr.ku.dk; soren.brunak@cpr.ku.dk
A list of authors and their ailiations appears at the end of the paper
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
had up to 24.7% missingness across the multi-omics data. For the clinical
data missingness was higher with a per individual median of 14% and 7%
for continuous and categorical clinical data, respectively. We designed
the MOVE framework to be flexible in relation to the number of input
data types and to be able to handle both continuous and categorical
features (Fig. 1a). To identify the optimal hyperparameters that would
capture the structure of the data without losing the ability to generalize
on unseen individuals, we initially divided the dataset into training and
test sets. We then measured the ability of the models to reconstruct the
input as well as the stability when refitting the model to the data several
times (Supplementary Figs. 2–4). The median reconstruction accura-
cies were between 0.95–1 and the final models were highly stable when
retrained five times with average change of cosine similarities in the
latent space of 0.037. Thus, the VAE models were able to reconstruct the
data with high accuracy across the individuals (Supplementary Fig. 5).
The latent space contains important clinical signatures
To illustrate how well the model captured the structure of the clinical
data, we analyzed the neural network weights connected to the input
variables of the encoder. Here we found the majority of the clinical
and dietary variables to be among the top 50 most important (Supple-
mentary Fig. 6). This was also the case when we investigated how the
continuous features impacted the positioning of the individuals in the
latent space using a Shapley additive explanation (SHAP) analysis
34
,
whereas for discrete features we found T2D-associated genetic variants
as well as clinically related features to be important (Supplementary
Fig. 7). Then, we investigated how individuals would be differenti-
ated by characteristics such as insulin sensitivity quantified by the
Matsuda index (Fig. 1b). Here we found a trend of the Matsuda Index
correlating with the two uniform manifold approximation and pro-
jection (UMAP) dimensions using Pearson’s correlation coefficient
(PCC) of 0.34 and −0.35 for dimensions one and two, respectively.
Using k-nearest-neighbor (kNN) regression on the latent representa-
tion we found that R2 for Matsuda Index (k = 5) was 0.70 compared to
0.37–0.38 when using residualized data or dimensionality reduction
using principal component analysis (PCA) and that this trend was con-
sistent for larger k (Supplementary Figs. 8 and 9). This indicated that
the MOVE latent representation captured a clinical signal that was
not as easily identified from the residualized data or by using PCA for
dimensionality reduction. Furthermore, we did not find any strong
local effects of missingness (R
2
= 0.05 at k = 5) and only small effects
of age (R2 < 0.01, k = 100). Similarly, we used a kNN classifier to inves-
tigate the effect of the confounders sex and recruitment center on the
global structure of the latent representation. These achieved accura-
cies of 0.58 and 0.25 for sex and center, respectively, which should
be compared to by-chance accuracies of 0.50 and 0.17, respectively
(Supplementary Figs. 10 and 11). If we used non-residualized data, that
is, when not correcting for confounding effects including age, sex, and
center, we observed larger effects (Supplementary Figs. 10 and 11).
This demonstrates the ability of the VAE to integrate heterogeneous
data but also that substantial confounding factors can influence the
latent representation.
Extracting drug to clinical and multi-omics associations
We then investigated if the model had learned associations between
the clinical, drug and multi-omics data. To do this, we developed an
approach that is based on perturbating input features one at a time
(Fig. 1a). For instance, to identify associations between a particular
drug and all other features, we simulated that we gave the drug to each
of the individuals that did not receive the drug. In addition to exclud-
ing individuals that were already receiving the drug we also excluded
individuals taking a drug of the same therapeutic drug-class in the
anatomical therapeutic chemical classification (ATC) system (Supple-
mentary Table 2). We then assessed if the change in each of the feature
reconstructions was significantly different compared to when passing
in mixOmics, which can integrate multiple modalities11,14–16. However,
these methods primarily focus on discovering factors or latent vari-
ables that can be used for visualization, clustering, or prediction of
disease.
We have previously developed a deep-learning framework on the
basis of variational autoencoders (VAE)
17,18
for integration and bin-
ning of large amounts of unstructured metagenomics data19. Specifi-
cally, a VAE is based on deep neural networks and learns to transform
high-dimensional data into a lower-dimensional space, termed a latent
representation. During this process the two networks of the VAE learn
the structure of input data and associations between the input vari-
ables. In our previous study, we found that the VAE could learn to inte-
grate two datasets without any prior knowledge or statistical model
19
.
Similarly, others have shown the capabilities of VAEs as integrative
models for extracting the underlying signal in data for improving clus-
tering and prediction
12,20–23
, as well as for handling large proportions of
missing data24. We, therefore, speculated that such a model could be
used to integrate even deeper cohort-level multi-omics datasets. While
previous studies have primarily focused on stratifying patients using
the underlying latent representation22,25,26 we were also interested in
whether we could acquire insights into the complex relationships that
the network learns through data integration.
For this purpose, we exploited that the decoder of the VAE is a
generative model. Thus, the final trained decoder will be able to gener-
ate new examples of data from the learned latent distribution. On the
basis of this principle, a variety of generative models have been used
to generate new examples of data, such as single-cell RNA data and
artificial human chromosomes
27,28
. Additionally, when combined with
Bayesian decision theory they have been used for analysis of single-cell
RNA data on the basis of variational inference29–31. Generative models
also allow investigation of the effect that a virtual perturbation of the
input data will have on the generated examples. For instance, Yeo et al.
trained a generative model on single-cell RNA time-series data and then
perturbed the input data to identify the effect of the perturbation on
the output of the generative model
32
. Similarly, a recent study used
the generative model of a VAE trained on protein evolutionary data to
predict the effect that genetic variants have on the fitness of human
proteins
33
. For our multi-modal data, we hypothesized that the genera-
tive ability of the VAE would allow us to identify associations between,
for example, patient exposures and omics features.
We therefore developed a framework that is based on VAEs that
we applied to a cohort of 789 people with newly diagnosed T2D with
extensive multi-omics characterization. These modalities included
genomics, transcriptomics, proteomics, metabolomics, and microbi-
omes as well as data on medication, diet questionnaires, and clinical
measurements. Our method was able to integrate multi-omics data with
clinical and categorical data and was resistant to systematic biases in
the data as well as large amounts of missing data. Using an ensemble
of generative VAE models, feature perturbation, univariate statistical
methods, and Bayesian decision theory we identify cross omics asso-
ciations. We compared the drug multi-omics profiles and showed
that different drugs are associated with unique clinical and molecular
profiles. Our method, multi-omics variational autoencoders (MOVE) is
freely available, easily scalable, can integrate any number of categorical
and continuous datasets, and able to identify features to multi-omics
associations.
Results
Designing a VAE for multi-omics data integration
We used a dataset of 789 newly diagnosed T2D individuals with exten-
sive multi-omics characterization (Supplementary Table 1). In total the
data included 8,807 variables per individual with median missingness
within an omics dataset of less than 5% except for metagenomics data
where two thirds of the individuals (532) did not have any data (Supple-
mentary Data 1 and Supplementary Fig. 1). Therefore, these individuals
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
the original data through the model (Fig. 1a). Because VAE models are
stochastic, we used results across an ensemble of models and devel-
oped two different approaches to identify significant associations. One
approach was based on applying t-tests with Bonferroni correction
across four different models, where each model was refitted 10 times
(MOVE t-test), while we also, inspired by earlier variational work
29–31
,
used Bayesian decision theory and a single model refitted 30 times
(MOVE Bayes). To identify different parameters of the approaches that
would allow for comparison across and to standard methods (t-test,
analysis of variance (ANOVA)), we applied them to two datasets con-
sisting of randomized clinical, drug and multi-omics data. Our find-
ings showed that MOVE t-test and MOVE Bayes had good performance
to identify drug–omics associations compared with t-test and ANOVA
at a ground-truth false discovery rate (FDR) of 0.05 (Supplementary
Fig. 12 and Supplementary Table 3 and Methods).
MOVE identifies drug and multi-omics associations
We then applied the MOVE framework to identify drug associations
in the DIRECT multi-modal data. The two methods, MOVE t-test and
MOVE Bayes, identified 3,143 and 763 significant associations to the
multi-omics and clinical features, respectively (Supplementary Tables
4–6 and Supplementary Data 2–4). We analyzed the intersection of the
two approaches and found that 573 of the 763 (75%) of the significant
associations were found by both methods (Fig. 1c). Making a conserva-
tive choice, we used the associations identified by both methods for fur-
ther analyses. When compared to traditional tests such as the Student’s
Trained VAE modelNaïve VAE model
Baseline model
Drug-perturbed model
Drug
0 1
Significant
drug–omics
associations
a
Multi-omics data
Non-omics data
Test
likelihood
VAE hyperparameter
selection
Model
accuracy
Model
stability
b
UMAP dimension 1
UMAP dimension 2
–1 0 1 2 3 4
6
8
9
10
5
7
Matsuda
–3 0 3
c
MOVE t-test
t-test
MOVE bayes
2,559
505
188
68
2
11
103
d
Acetylsalicylic acid
Amlodipine
Atenolol
Atorvastatin
Bendroflumethiazide
Bisoprolol
Codeine
Enalapril
Hydrochlorothiazide
Lansoprazole
Levothyroxine sodium
Lisinopril
Losartan
Metformin
Metoprolol
Omeprazole
Paracetamol
Ramipril
Salbutamol
Simvastatin
0
25
50
75
Drug–multi-omics associations
t−test
ANOVA
MOVE overlap
e
Clinical continuous
Diet and wearables
Proteomics
Targeted metabolomics
Untargeted metabolomics
Transcriptomics
Metagenomics
0
0.05
0.10
0.15
0.20
Fraction significant drug
associations
Fig. 1 | Integrating multi-omics data with a VAE. a, Principle of integration
and analysis approach using MOVE. Individual-level non-omics and multi-omics
data were used as input to a VAE. The optimal network hyperparameters were
estimated from the summed test set error across all individuals in the test (test
likelihood), training reconstruction accuracy, and model stability. Significant
drug–omics associations were identified by perturbing drug status from no (0)
to yes (1) for all individuals that were not already administered the drug. b, UMAP
representation of the latent representation from the 789 people with newly
diagnosed T2D. Individuals were colored according to their z-scaled Matsuda
index from low (blue), average (yellow), and high (red). c, Overlap in significant
drug–omics associations between standard t-test (two-sided, Benjamini–
Hochberg FDR < 0.01) on the input data, MOVE t-test (multi-stage Bonferroni-
corrected, P adjust < 0.05) and MOVE Bayes approaches (FDR Bayes < 0.05). The
different methods of multiple testing correction corresponded to FDR of 0.05
on the ground-truth dataset. The overlap between MOVE t-test and MOVE Bayes
was used for further analysis (n = 573). d, The number of significant associations
found between drugs and features in the multi-omics datasets using MOVE t-test
and MOVE Bayes (purple), t-test (green) or ANOVA (orange). See c for information
on the tests. e, Fraction of features in the multi-omics datasets that was found by
MOVE to be significantly associated with at least one drug (n = 20). The lower and
upper hinges correspond to the first and third quartiles. The upper and lower
whiskers extend from the hinge to the highest and lowest values, respectively, but
no further than 1.5× interquartile range from the hinge. Data beyond the ends of
whiskers are outliers and are plotted individually.
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
t-test and ANOVA we found this to add 211% more significant associa-
tions, from 184 to 573 (Fig. 1d). In addition, the significant associations
identified by MOVE were distributed across the drugs (two-sided t-test,
P = 0.016) and not only for the drugs administered to most individuals
such as Simvastatin, Atorvastatin, and Metformin. For instance, MOVE
identified a median of 20 associations per drug compared to 1 for t-test
and 0 for ANOVA, highlighting that our method was more sensitive for
extracting associations for drugs given to a smaller number of individu-
als (Supplementary Tables 5 and 6). Among the multi-omics datasets,
we found that the largest number of significant drug associations was
to the metabolomics, clinical, and transcriptomics data with an aver-
age of six associations per drug (Fig. 1e and Supplementary Fig. 13).
When normalizing for all possible associations, the highest fraction
of asso ciations was to the clinical data (8%) followed by targeted
and untargeted metabolomics with an average of 5.1% and 2.8% of the
features associated to a drug, respectively. Finally, we investigated if
our results could be driven by disease subtypes within the T2D cohort.
To do this, we used four archetype clusters from Wesolowska–Andersen
and Brorsson et al.7 that were based on clustering from 32 clinical
features. Here we found that a median of 6.5% of the significant drug–
omics associations were specific to one of the subgroups indicating
that the associations were not primarily driven by the archetypes
(Supplementary Table 7).
Changes in T2D biomarkers were associated with metformin
We then investigated drug and multi-omics interactions (Fig. 2a and
Supplementary Figs. 14–18), and initially focused on expected clinical
drug interactions. For instance, for metformin, we identified 88 sig-
nificant clinical and multi-omics interactions across all the datasets.
When investigating associations across the individuals we found low
intra-patient variability indicating that the changes were stable (Fig. 2b
and Supplementary Fig. 19). We found that metformin was significantly
associated with 12 clinical markers of T2D such as insulin clearance,
active GLP-1, glucose levels from mixed-meal glucose tolerance test,
glucose sensitivity, and blood pressure (Fig. 2a and Supplementary
Data 2–4). The directions of some of the associations were opposite
to the expected metformin effects, that is, metformin was associated
with decreased glucose sensitivity at baseline (average Z-score change
Metformin–multi-omics features
Individuals
Eect size (z-scaled)
Drugs
Significant drug–clinical associations
ab
Glu.sens
hsCRP
Fasting.LDL
Fasting.Chol
Fasting.GAD
Fasting.UCpep
Fasting.UCreatinine
WHR
Fasting.Creatinine
BSA
Rate.sens
Fasting.UCPCR
Fasting.HDL
Active_glp1_conc_0 min
Glucagon_conc_60 min_pg_ml
Total_glp1_conc_0 min
Total_glp1_conc_60 min
Fasting.ALT
Fasting.AST
mmtt.120.Glucose
Mean.glu
Energy intake
bmi
Fasting.TG
mmtt.120.Insulin
mmtt.120.Cpep
HOMA2..B
Mean.ins
Fasting.Insulin
Basal.ins
Liver.iron
Fasting.GADaab
Asat
Liver.fat
Iaat
Panc.fat
Fasting.HbA1c
Proinsulin_conc_60min_pmol_L
Stumvoll
Clins
Matsuda
Clinsb
PFR1
BP_s_mean
BP_d_mean
* * * * *
* * * * * *
**** * *
*** * * * *
* * ** * * *
* *
** * * * *
**
* * *
* * * *
* * * *
**
* * *
***
* * *
* *
* * *
**
**
** *
* *
* * * *
**
*
* *
* *
* *
* * *
* * *
* * *
**
* *
*
*
* * * *
*
* * *
** *
**
** * *
−0.03
−0.02
−0.01
0
0.01
0.02
Metformin
Levothyroxine sodium
Simvastatin
Acetylsalicylic acid
Ramipril
Omeprazole
Losartan
Atorvastatin
Enalapril
Atenolol
Lansoprazole
Lisinopril
Hydrochlorothiazide
Metoprolol
Bendroflumethiazide
Salbutamol
Paracetamol
Codeine
Amlodipine
Bisoprolol
CPT1A
33442
32746
32393
17945
1121
41754
Pro
1898
IRX2
CHOm.g
TFF3_pro
CD40LG
Fasting.UCpep
Selenium.ug
Fasting.GAD
ALAS2
UBB
CD177
43258
32620
Rate.sens
Mean.glu
Clinsb
Clins
Active_glp1_conc_0 min
mmtt.120.Glucose
MGS:igc1047
MGS:igc0971
MGS:igc0313
MGS:igc0237
MGS:igc0295
CXCL8
MPPE1
CEP19
ICAM1
CDKN1C
TCF7L2
RASA4
RASA4B
NID1
PSMA1
32587
569
18394
32497
40135
32759
39517
32839
32847
32616
C8
33968
1642
ERAP2
CCDC151
BP_s_mean
MGS:igc0006
MGS:igc1492
MGS:igc0392
MGS:igc0029
MGS:igc0655
MGS:igc0359
MGS:igc0985
MGS:igc0641
MGS:igc0658
Arg
Trp
Tyr
SSTR3
PC.ae.C30.0
PC.ae.C32.1
PC.ae.C32.2
PC.ae.C34.2
PC.ae.C36.3
PC.ae.C36.5
PC.ae.C36.4
PC.ae.C38.5
PC.aa.C36.0
PC.aa.C38.0
PC.ae.C38.6
PFR1
mean.ins
Glu.sens
36079
C22orf24
C0
−0.04
−0.02
0
0.02
0.04
0.06
Eect size (z-scaled)
Fig. 2 | Significant associations between drugs, clinical, and multi-omics
features. a, Significant associations between drugs and clinical features.
Effects are given as effect size (z-scaled units) from negative (blue) to positive
(red). Significant associations identified by both MOVE t-test and MOVE Bayes
are indicated using a star. Features (y-axis) and drugs (x-axis) are clustered
using hierarchical clustering on the basis of Euclidean distances. b, As in a but
showing per individual-level associations of metformin to multi-omics features
demonstrating that associations are highly stable across individuals. Features
(y-axis) and newly diagnosed T2D individuals (x-axis).
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
−0.029, confidence intervals [−0.030, −0.029]). This could be due to
confounding by indication in terms of the study design where newly
diagnosed T2D individuals that have been prescribed metformin
are expected to have more severe clinical T2D values compared to
individuals not needing medical treatments
35,36
. Therefore, since all
individuals have T2D the confounding effect of their diabetic status
could not be disentangled from the effect of metformin. When inves-
tigating the multi-omics associations of metformin we found two of
the seven associated proteins (ERAP2 and CD40L) could be linked to
the immune system (Fig. 3a and Supplementary Data 4). Similarly, for
the transcriptomics data we found CXCL8 and CD177 to be altered by
metformin where the former has been shown to be altered in healthy
individuals and cancer patients
37–39
. In the targeted metabolomics data
we identified a significant enrichment of metabolites associated with
aminoacyl-tRNA biosynthesis (hypergeometric test, P = 2.2 × 10−4, FDR
corrected). This pathway has previously been associated with met-
formin in functional pathway analysis of microbial change in mice
40
.
Finally, for the untargeted metabolomics data, metformin had the
highest number of associations of any drug (22 associations) indicating
that new metabolic effectors of metformin treatment could potentially
be identified (Supplementary Fig. 17 and Supplementary Table 4).
Association of metformin and omeprazole with gut
microbiota
Recent studies have shown how drug intake can influence the human
gut microbiome composition41,42. Here we found metformin and
Bacteroidales sp. igc0029
Faecalibacterium prausnitzii igc0359
Anaerostipes sp. igc0655
S. parasanguinis igc0686
Clostridia sp. igc0658
Bacteroidetes sp. igc0392
Escherichia coli igc0006
Betaproteobacteria sp. igc0985
Parasutterella sp. igc1492
S. vestibularis igc1414
Streptococcus sp. igc0014
Clostridiales sp. igc0641
Clostridiaceae sp. igc1047
Clostridiales sp. igc0971
Intestinibacter bartlettii igc0313
Peptostreptococcaceae sp. igc0295
Romboutsia timonensis igc0237
Drug
Metformin
Omeprazole
Effect size (z−scale)
−0.01 0 0.01
Salbutamol
Lansoprazole
Paracetamol
Codeine
Simvastatin
Bendroflumethiazide
Lisinopril
Levothyroxine sodium
Enalapril
Atenolol
Metformin
Losartan
Hydrochlorothiazide
Metoprolol
Omeprazole
Ramipril
Acetylsalicylic acid
Bisoprolol
Amlodipine
Atorvastatin
Salbutamol
Lansoprazole
Paracetamol
Codeine
Simvastatin
Bendroflumethiazide
Lisinopril
Levothyroxine Sodium
Enalapril
Atenolol
Metformin
Losartan
Hydrochlorothiazide
Metoprolol
Omeprazole
Ramipril
Acetylsalicylic acid
Bisoprolol
Amlodipine
Atorvastatin
0
1
Cosine
similarity
Average drug omics effect size
0.0025
0.0050
0.0075
0.0100
Metformin
Atorvastatin
Simvastatin
Omeprazole Metformin
Atorvastatin
Simvastatin
Omeprazole
Metformin
Atorvastatin
Simvastatin
Omeprazole
Metformin
Atorvastatin
Simvastatin
Omeprazole
Metformin
Atorvastatin
Simvastatin
Omeprazole
Metformin
Atorvastatin
Simvastatin
Omeprazole
Metformin
Atorvastatin
Simvastatin
Omeprazole
Clinical continuous
Diet and wearables
Proteomics
Targeted metabolomics
Unargeted metabolomics
Transcriptomics
Metagenomics
Metformin
Omeprazole
Acetylsalicylic acid
Amlodipine
Bisoprolol
Paracetamol
Bendroflumethiazide
Hydrochlorothiazide
Ramipril
Levothyroxine sodium
Losartan
Lansoprazole
Metoprolol
Atorvastatin
Atenolol
Codeine
Salbutamol
Lisinopril
Enalapril
Simvastatin
b c
d
Streptococcus parasanguinis
F. prausnitzii
Streptococcus vestibularis
Intestinibacter bartlettii
Streptococcus sp.
Escherichia coli
Cholesterol
homeostasis
Lipid homeostasis
Active glp1 0 min
Fasting urine creatinine
MMT 120 glucose
Fasting HDL
Fasting cholesterol
Fasting LDL
Glucose sens.
Diurnal rmse
Fibre AOAC
Selenium
Carotene
Transfats
CHOm
EIF2AK3
CD40LG
IGFBP4
IGFBP1
ERAP2
FADS1
Gly
Pro
Trp
Tyr
Arg
SREBF2
ABCG1
ABCA1
MYLIP
CXCL8
LDLR
Clinical
Diet and wearables
Proteomics
Metabolomics
Untargeted
metabolomics
Transcriptomics
Metagenomics
Gene Ontologies
Metformin
Simvastatin
Atorvastatin
Omeprazole
Lansoprazole
Paracetamol
Codeine
Impact
Drug ~ omics
associations
a
Drug effect rank in multi-omics data
5
10
15
20
100
200
300
Number of individuals
e
Fig. 3 | Drug associations with metagenomics species and drug–drug
similarities. a, Display of effect sizes (z-scaled units) for (outer to inner)
metformin, simvastatin, atorvastatin, omeprazole, lansoprazole, paracetamol,
and codeine. Only significant associations to any of the drugs are shown and
effect size is visualized as brown (negative), gray (none), and green (positive).
Selected omics features are indicated. The Gene Ontologies element represents
significantly over-represented Gene Ontology terms using transcriptomics
(hypergeometric test, FDR < 0.05) (green). The innermost ring indicates SHAP
importance for the individual features in the encoding from input data to the
latent representation. b, Effect size (z-scaled units) (x-axis) of the human gut
metagenomics species that were significantly associated with metformin
(orange) or omeprazole (teal). c, Drug–drug similarities by comparing
drug-response profiles across the multi-omics datasets. Cosine similarity
indicated from no similarity (blue) to identical profiles (red). d, Average effect
(z-score) of drugs for the omics datasets. All 20 drugs are shown, however, only
metformin (red), omeprazole (purple), atorvastatin (green), and simvastatin
(blue) are indicated. All other drugs are colored gray without a text label.
e, Distribution of multi-omics ranks for the different drugs. The ranks are
determined as a number between 1–20 (drugs) on the basis of the average effect
size from d. The boxes are colored according to number of individuals taking a
particular drug from 0 (white) to 323 (purple). There was no correlation between
rank scores and number of individuals taking a drug (PCC = 0.14). The lower and
upper hinges correspond to the first and third quartiles. The upper and lower
whiskers extend from the hinge to the highest and lowest values, respectively, but
no further than 1.5× interquartile range from the hinge. Data beyond the ends of
whiskers are outliers and are plotted individually.
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
omeprazole to be the only drugs to have significant associations
to the metagenomics data with an increase of eleven metagenomics
species as well as a decrease of six other species (Fig. 3b). Remark-
ably, the findings of increased Escherichia coli and decreased levels
of Intestinibacter bartlettii and Peptostreptococcaceae sp. have been
reported in healthy individuals taking metformin in an intervention
study43 (Supplementary Data 4). As the study first reporting the findings
was performed in healthy individuals, the changes are most likely not
explained by other factors than metformin treatment. For omepra-
zole, a protein pump inhibitor (PPI), we identified three Streptococcus
species to be significantly increased (Streptococcus sp., Streptococcus
parasanguinis, and Streptococcus vestibularis) (Supplementary Data
4). Previous work by others has specifically shown PPIs to influence
the abundance of Streptococcus parasanguinis and vestibularis in
the human gut
44
. Interestingly, both omeprazole and lansoprazole
target the K-transporter ATPase alpha channel 1 and increases pH in
the stomach. The two drugs, however, have different speed to effect
rates where omeprazole elicits its effect with a slower rate compared
to lansoprazole
45
. This, in combination with more individuals being
administered omeprazole (125) compared to lansoprazole (57), could
explain why we identified significant alterations of gut microbiota for
omeprazole and not lansoprazole.
Statins were associated with decreased low-density
lipoprotein and cholesterol
Next, we investigated associations between the two statins, simvastatin,
and atorvastatin, which are widely used to treat high blood choles-
terol by lowering low-density lipoprotein (LDL)46. In agreement with
their potential to treat dyslipidemia, we found both LDL and overall
cholesterol levels to be significantly associated and decreased with
average LDL z-score change of −0.039 (CI [−0.040, −0.038]) and −0.015
(CI [−0.016, −0.014]) for simvastatin and atorvastatin, respectively
(Supplementary Data 4). This effect could be a consequence of many
of the participants having been administered statins before their T2D
diagnosis (simvastatin median duration 1.9 years and atorvastatin
median duration 1.7 years; Supplementary Table 8), thereby increasing
the chance of observing the effect of the drug with reduced confound-
ing by indication. Interestingly, we noticed that besides the down-
regulation of LDL and general cholesterol levels some of the remaining
clinical associations were not similar. Simvastatin was associated with
an increase in the health marker high-density lipoprotein (HDL) cho-
lesterol whereas atorvastatin had a decrease. This agrees with known
effects of the two statins on HDL, where simvastatin and atorvastatin,
respectively, increase and decrease HDL levels with increasing doses
47
.
Different molecular profiles of simvastatin and atorvastatin
When investigating the multi-omics associations, the two statins had
diverse effects across the omics data (Fig. 3a and Supplementary Figs.
14–18 and 20). In agreement with the analysis of the clinical data, we
found simvastatin to be significantly associated with downregulation
of cholesterol homeostasis (Hypergeometric test, P = 0.005, FDR) and
lipid transportation pathways (Hypergeometric test, P = 0.002, FDR)
from the enrichment analysis of the associated transcripts (Fig. 3a
and Supplementary Data 4 and 5). Specifically, we identified changes
in LDLR, SREBF2, ABCA1, and ABCG1 expression, previously associated
with simvastatin usage and accumulation of fatty acid and triglyceride
in the liver through different pathways
48–52
(Supplementary Data 4). In
the proteomics data of atorvastatin, we identified known associations
to FADS1 (ref.
53
), as well as EIF2AK3, which has been reported associ-
ated with cholesterol homeostasis
54,55
. Additionally, two insulin growth
factor binding proteins (IGFBP1 and IGFBP4) were associated with ator-
vastatin and IGFBP4 for simvastatin as well (Supplementary Data 4).
These have previously been reported specifically for people with T2D
and atorvastatin use
54,56
. Finally, in the targeted metabolomics data,
we identified simvastatin to be associated with an increase in glycine
levels, which in low systemic concentration has been associated with
obesity and T2D
57
(Supplementary Data 4). Furthermore, we observed
a decrease of several phosphatidylcholines (11 of 17 decreased metabo-
lites), and an increase of sphingomyelin and ceramide (2 of 11 increased
metabolites), a ratio which has previously been shown to be altered with
high doses of simvastatin compared to other statins
58
(Supplementary
Data 2–4). For atorvastatin, we observed a non-significant decrease of
glycine levels and that the overall ratio of sphingomyelin and ceramide
decreased (4 of 13 decreased metabolites).
Drug polypharmacy and similarity across multi-omics data
We then investigated similarities between drugs and their multi-omics
associations. Overall, we observed four clusters containing three to
six drugs each and found that some of the drugs within a cluster could
potentially be associated with polypharmacy (Fig. 3c). Therefore, we
investigated the impact of a drug–drug combination on the associa-
tions and found a correlation between overall drug association simi-
larity and the individuals taking the two drugs (PCC 0.75, P value of
2.2 × 10
−35
). This finding indicates possible polypharmacy effects intro-
duced by taking the two drugs together resulting in a higher drug–drug
similarity across all clinical and multi-omics changes. However, some of
the similarities might to some extent be driven by overlapping patient
groups and non-drug-related similarities such as the underlying reason
for taking the drug. An example could be the drug similarity cluster of
Ramipril, Acetylsalicylic Acid, Bisoprolol, Amlodipine and Atorvasta-
tin, which can be linked to cardiovascular diseases. Furthermore, the
drugs that had the most similar drug and multi-omics associations were
codeine and paracetamol with a cosine similarity of 0.78. Most (38 of 46)
of the individuals in the cohort taking codeine were also taking paracet-
amol while a large fraction of individuals (52 of 90) was only taking par-
acetamol. We therefore cannot rule out that the correlated multi-omics
profiles of the two drugs could be driven by the partial overlap leading
to similar latent representation and model reconstructions. Finally,
we investigated known drug–drug interactions and association with
drug multi-omics profiles; however, found no statistically significant
correlations (Supplementary Note and Supplementary Fig. 21).
The effects of drugs are widespread across the omics data
Currently, there are widespread efforts in investigating drugs and gut
microbiome interactions suggesting that the microbiome is a potential
target and mediator of drug effect
42,59,60
. As we investigated several
multi-omics datasets besides the gut microbiome (metagenomics),
we can compare the effect size of the drugs across the omics datasets.
Interestingly, we found that the gut microbiome was the dataset with
the second fewest number of statistically significant hits across the
drugs with 17 significant associations (Supplementary Table 4 and
Supplementary Fig. 13). Only diet and wearable data had fewer asso-
ciations (11); transcriptomics, proteomics, targeted, and untargeted
metabolomics had between 44–134 significant associations. We then
asked if the effect size of the drugs were different across datasets and
determined the cumulative effect size of the drugs in the respective
multi-omics datasets. Here we found that the average effect sizes in
transcriptomics and metagenomics data were the lowest for all drugs,
and that those in the metagenomics dataset were significantly lower
compared to all other omics datasets but transcriptomics (ANOVA,
Tukey HSD test, adjusted P < 0.05) (Fig. 3d and Supplementary Table 9).
When we subset to significant drug–omics associations, of which
the gut microbiome only had two drugs with significant associations
(metformin and omeprazole), we found that the effect of these two
drugs were similar or lower compared to the effect sizes of the other
multi-omics datasets (Supplementary Fig. 22). Finally, we investigated
if this could be caused by increased uncertainty when learning and
reconstructing a given modality but only found small correlations with
PCCs of −0.15 to 0.16 between modality uncertainty and inferred effect
sizes in a modality (Supplementary Table 10). Overall, this observation
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
implies that the multi-omics response to drug stimuli are not only
targeting the gut microbiome and that multiple omics datasets should
be included when attempting to understand drug effects.
Ranking the impact of drugs in multi-omics data
Finally, we investigated the effect sizes of the individual drugs across
the multi-omics datasets. We found that metformin and omeprazole,
in general, had the most pronounced effects on the multi-omics data
(cumulative rank scores) and that the two statins ranked 14 and 20 out
of the 20 drugs (Fig. 3e) where simvastatin had the lowest overall rank
of cumulative effect sizes. This analysis was not confounded by the
number of individuals taking a particular drug as there was no correla-
tion (PCC = 0.14) between the number of individuals and drug effect.
This was opposed to when investigating only significant associations
where statins ranked 2 and 4 with high effect sizes (Supplementary
Figs. 22 and 23). This observation may indicate that statins had fewer
strong effects, whereas, for instance, both metformin and omeprazole
with the highest average rank had larger systemic effects.
Discussion
Here we show that it is possible to use unsupervised deep learning to
integrate and extract associations from a deeply phenotyped cohort
of people with T2D. While existing methods for vertical integration
of multi-omics data focus on encoding the data to factors or latent
representations that can be used for clustering and classification, we
took this further by using the generative capacity of VAE models. In
comparison to traditional univariate statistical tests, MOVE can identify
significant drug–omics associations for a wider selection of drugs. We
believe that these improvements come from the ability of the genera-
tive models to infer multi-omics changes for individuals not receiving
a drug thus increasing power.
Previous work to stratify the newly diagnosed T2D individuals
from this cohort used 32 clinical features to identify four archetypes
representing different T2D subtypes7. In addition, they used met-
formin status of the individuals to investigate if the subgroups were
confounded by metformin treatment and found no significant impact
on the clusters and their multi-omics correlations. In contrast to their
work, we added medication data on 19 additional drugs and used all
data as input to our unsupervised deep-learning model allowing the
model to learn from all inputs simultaneously. Thus, we were able to
identify associations between the drugs and multi-omics data, includ-
ing for metformin indicating the importance of vertical integration.
The cross-sectional design and clinical data-guided medical deci-
sions make it difficult to assess the directionality of drug associations
and further complicates causal inference. Hence, it is not possible to
draw causal conclusions on drug effects; however, the results can be
considered as input to design informed studies as well as randomized
clinical control studies. In the future, expansion with longitudinal
multi-omics data and modeling time could add more information on
the causality of the drugs by investigating the long-term effects and
associations32.
Similarly, our approach opens up for individualized analysis of
patients in an N-of-1 approach61. It is well-known in health care that often
selecting a drug or treatment in a situation at the same time excludes
performing the control experiment of using another drug. Using MOVE,
we can in principle ask what would happen if we gave the patient a drug
and compare to the result of choosing another drug. Our cohort size
is limited, but for larger cohorts of tens to hundreds of thousands of
patients this could potentially be powerful to identify molecular asso-
ciations and treatment outcomes for individual patients.
Finally, we emphasize that our approach is, of course, not limited
to drug associations; in principle, all the omics data could be assessed
for associations across the datasets. We therefore believe that our
generative method opens new possibilities in big multi-omics data
analysis for discoveries of potential new biomarkers, carrying out
gedankenexperiments, and investigating potential direct effects of
drugs in high dimensionality molecular data that leads to testable
hypotheses.
Online content
Any methods, additional references, Nature Research reporting sum-
maries, source data, extended data, supplementary information,
acknowledgements, peer review information; details of author contri-
butions and competing interests; and statements of data and code avail-
ability are available at https://doi.org/10.1038/s41587-022-01520-x.
References
1. Fares, H., DiNicolantonio, J. J., O’Keefe, J. H. & Lavie, C. J.
Amlodipine in hypertension: a irst-line agent with eicacy for
improving blood pressure and patient outcomes. Open Heart 3,
e000473 (2016).
2. Hu, J. X., Thomas, C. E. & Brunak, S. Network biology concepts
in complex disease comorbidities. Nat. Rev. Genet. 17, 615–629
(2016).
3. Austin, R. P. Polypharmacy as a risk factor in the treatment of
type 2 diabetes. Diabetes Spectr. 19, 13–16 (2006).
4. Zhou, W. et al. Longitudinal multi-omics of host–microbe
dynamics in prediabetes. Nature 569, 663–671 (2019).
5. Hasin, Y., Seldin, M. & Lusis, A. Multi-omics approaches to disease.
Genome Biol. 18, 83 (2017).
6. Gudmundsdottir, V. et al. Whole blood co-expression modules
associate with metabolic traits and type 2 diabetes: an IMI-DIRECT
study. Genome Med. 12, 109 (2020).
7. Wesolowska-Andersen, A. et al. Four groups of type 2 diabetes
contribute to the etiological and clinical heterogeneity in newly
diagnosed individuals: an IMI DIRECT study. Cell Reports Medicine
3, 100477 (2022).
8. Song, J. W. & Chung, K. C. Observational studies: cohort and
case-control studies. Plast. Reconstr. Surg. 126, 2234–2242 (2010).
9. Picard, M., Scott-Boyer, M.-P., Bodein, A., Périn, O. & Droit, A.
Integration strategies of multi-omics data for machine learning
analysis. Comput. Struct. Biotechnol. J. 19, 3735–3746 (2021).
10. Nicora, G., Vitali, F., Dagliati, A., Geifman, N. & Bellazzi, R.
Integrated multi-omics analyses in oncology: a review of machine
learning methods and tools. Front. Oncol. 10, 1030 (2020).
11. Rohart, F., Gautier, B., Singh, A. & Lê Cao, K.-A. mixOmics:
an R package for ’omics feature selection and multiple data
integration. PLoS Comput. Biol. 13, e1005752 (2017).
12. Chung, N. C. et al. Unsupervised classiication of multi-omics data
during cardiac remodeling using deep learning. Methods 166,
66–73 (2019).
13. Kriebel, A. R. & Welch, J. D. UINMF performs mosaic integration
of single-cell multi-omic datasets using nonnegative matrix
factorization. Nat. Commun. 13, 780 (2022).
14. Argelaguet, R. et al. Multi-omics factor analysis—a framework for
unsupervised integration of multi-omics data sets. Mol. Syst. Biol.
14, e8124 (2018).
15. Shen, R., Olshen, A. B. & Ladanyi, M. Integrative clustering of
multiple genomic data types using a joint latent variable model
with application to breast and lung cancer subtype analysis.
Bioinformatics 25, 2906–2912 (2009).
16. Singh, A. et al. DIABLO: an integrative approach for identifying
key molecular drivers from multi-omics assays. Bioinformatics 35,
3055–3062 (2019).
17. Kingma, D. P. & Welling, M. Auto-Encoding Variational Bayes.
Preprint at arXiv https://doi.org/10.48550/arXiv.1312.6114 (2013).
18. Rezende, D. J., Mohamed, S. & Wierstra, D. Stochastic
backpropagation and approximate inference in deep generative
models. Preprint at arXiv https://doi.org/10.48550/arXiv.1401.4082
(2014).
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
19. Nissen, J. N. et al. Improved metagenome binning and assembly
using deep variational autoencoders. Nat. Biotechnol. 39,
555–560 (2021).
20. Ding, J., Condon, A. & Shah, S. P. Interpretable dimensionality
reduction of single cell transcriptome data with deep generative
models. Nat. Commun. 9, 2002 (2018).
21. Chaudhary, K., Poirion, O. B., Lu, L. & Garmire, L. X. Deep
learning-based multi-omics integration robustly predicts survival
in liver cancer. Clin. Cancer Res. 24, 1248–1259 (2018).
22. Zhang, L. et al. Deep learning-based multi-omics data integration
reveals two prognostic subtypes in high-risk neuroblastoma.
Front. Genet. 9, 477 (2018).
23. Cao, Z. -J. & Gao, G. Multi-omics single-cell data integration
and regulatory inference with graph-linked embedding.
Nat. Biotechnol. 40, 1458–1466 (2022).
24. Mattei, P.-A. & Frellsen, J. MIWAE: deep generative modelling
and imputation of incomplete data. In Proceedings of the 36th
Interna tional Conference on Machine Learning 4413–4423
(PMLR, 2019).
25. Way, G. P. & Greene, C. S. Extracting a biologically relevant latent
space from cancer transcriptomes with variational autoencoders.
Pac. Symp. Biocomput. 23, 80–91 (2018).
26. Allesøe, R. L. et al. Deep learning-based integration of genetics
with registry data for stratiication of schizophrenia and
depression. Sci. Adv. 8, eabi7293 (2022).
27. Ghahramani, A., Watt, F. M. & Luscombe, N. M. Generative
adversarial networks simulate gene expression and predict
perturbations in single cells. Preprint at bioRxiv https://doi.org/
10.1101/262501 (2018).
28. Yelmen, B. et al. Creating artiicial human genomes using
generative neural networks. PLoS Genet. 17, e1009303 (2021).
29. Lopez, R., Regier, J., Cole, M. B., Jordan, M. I. & Yosef, N. Deep
generative modeling for single-cell transcriptomics. Nat. Methods
15, 1053–1058 (2018).
30. Gayoso, A. et al. A Python library for probabilistic analysis of
single-cell omics data. Nat. Biotechnol. 40, 163–166 (2022).
31. Lopez, R., Boyeau, P., Yosef, N., Jordan, M. I. & Regier, J. Decision-
making with auto-encoding variational Bayes. In Proceedings
of the 34th International Conference on Neural Information
Processing Systems 5081–5092 (Curran Associates Inc., 2020).
32. Yeo, G. H. T., Saksena, S. D. & Giord, D. K. Generative modeling
of single-cell time series with PRESCIENT enables prediction
of cell trajectories with interventions. Nat. Commun. 12, 3222
(2021).
33. Frazer, J. et al. Disease variant prediction with deep generative
models of evolutionary data. Nature 599, 91–95 (2021).
34. Lundberg, S. M. & Lee, S.-I. A uniied approach to interpreting
model predictions. in Advances in Neural Information Processing
Systems 30 (eds Guyon, I. et al.) 4765–4774 (Curran Associates,
2017).
35. Hirst, J. A., Farmer, A. J., Ali, R., Roberts, N. W. & Stevens, R. J.
Quantifying the eect of metformin treatment and dose on
glycemic control. Diabetes Care 35, 446–454 (2012).
36. Knowler, W. C. et al. Reduction in the incidence of type 2 diabetes
with lifestyle intervention or metformin. N. Engl. J. Med. 346,
393–403 (2002).
37. Ustinova, M. et al. Metformin strongly aects transcriptome
of peripheral blood cells in healthy individuals. PLoS One 14,
e0224835 (2019).
38. Xiao, Z., Wu, W. & Poltoratsky, V. Metformin suppressed CXCL8
expression and cell migration in HEK293/TLR4 cell line. Mediators
Inlamm. 2017, 6589423 (2017).
39. Bruno, S. et al. Metformin inhibits cell cycle progression of B-cell
chronic lymphocytic leukemia cells. Oncotarget 6, 22624–22640
(2015).
40. Ma, W. et al. Metformin alters gut microbiota of healthy mice:
implication for its potential role in gut microbiota homeostasis.
Front. Microbiol. 9, 1336 (2018).
41. Forslund, K. et al. Disentangling type 2 diabetes and metformin
treatment signatures in the human gut microbiota. Nature 528,
262–266 (2015).
42. Vieira-Silva, S. et al. Statin therapy is associated with lower
preva lence of gut microbiota dysbiosis. Nature 581, 310–315
(2020).
43. Bryrup, T. et al. Metformin-induced changes of the gut microbiota
in healthy young men: results of a non-blinded, one-armed
intervention study. Diabetologia 62, 1024–1035 (2019).
44. Vich Vila, A. et al. Impact of commonly used drugs on the
composition and metabolic function of the gut microbiota.
Nat. Commun. 11, 362 (2020).
45. Shin, J. M., Munson, K., Vagin, O. & Sachs, G. The gastric
HK-ATPase: structure, function, and inhibition. Plugers Arch. 457,
609–622 (2009).
46. Cholesterol Treatment Trialists’ (CTT) Collaboration. et al. Eicacy
and safety of more intensive lowering of LDL cholesterol: a
meta-analysis of data from 170,000 participants in 26 randomised
trials. Lancet 376, 1670–1681 (2010).
47. Barter, P. J., Brandrup-Wognsen, G., Palmer, M. K. & Nicholls, S. J.
Eect of statins on HDL-C: a complex process unrelated to
changes in LDL-C: analysis of the VOYAGER database. J. Lipid Res.
51, 1546–1553 (2010).
48. Aguayo-Orozco, A. et al. sAOP: linking chemical stressors
to adverse outcomes pathway networks. Bioinformatics 35,
5391–5392 (2019).
49. Margerie, D. et al. Hepatic transcriptomic signatures of statin
treatment are associated with impaired glucose homeostasis in
severely obese patients. BMC Med. Genomics 12, 80 (2019).
50. Gilbert, R., Al-Janabi, A., Tomkins-Netzer, O. & Lightman, S. Statins
as anti-inlammatory agents: a potential therapeutic role in sight-
threatening non-infectious uveitis. Porto Biomed J 2, 33–39 (2017).
51. Aguayo-Orozco, A., Bois, F. Y., Brunak, S. & Taboureau, O. Analysis
of time-series gene expression data to explore mechanisms of
chemical-induced hepatic steatosis toxicity. Front. Genet. 9, 396
(2018).
52. Kennedy, M. A. et al. ABCG1 has a critical role in mediating
cholesterol elux to HDL and preventing cellular lipid
accumulation. Cell Metab. 1, 121–131 (2005).
53. Ishihara, N. et al. Atorvastatin increases Fads1, Fads2
and Elovl5 gene expression via the geranylgeranyl
pyrophosphate-dependent Rho kinase pathway in 3T3-L1 cells.
Mol. Med. Rep. 16, 4756–4762 (2017).
54. Ferretti, G., Bacchetti, T., Banach, M., Simental-Mendía, L. E. &
Sahebkar, A. Impact of statin therapy on plasma MMP-3, MMP-9,
and TIMP-1 concentrations: a systematic review and meta-analysis
of randomized placebo-controlled trials. Angiology 68, 850–862
(2017).
55. Orekhov, A. N. et al. Role of phagocytosis in the pro-inlammatory
response in LDL-induced foam cell formation; a transcriptome
analysis. Int. J. Mol. Sci. 21, 817 (2020).
56. Osório, J. Statins and T2DM—an IGF link? Nat. Rev. Endocrinol. 9,
187–187 (2013).
57. Alves, A., Bassot, A., Bulteau, A.-L., Pirola, L. & Morio, B. Glycine
metabolism and its alterations in obesity and metabolic diseases.
Nutrients 11, 1356 (2019).
58. Snowden, S. G. et al. High-dose simvastatin exhibits enhanced
lipid-lowering eects relative to simvastatin/ezetimibe
combination therapy. Circ. Cardiovasc. Genet. 7, 955–964
(2014).
59. Forslund, S. K. et al. Combinatorial, additive and dose-dependent
drug-microbiome associations. Nature 600, 500–505 (2021).
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
60. Zimmermann, M., Zimmermann-Kogadeeva, M., Wegmann, R. &
Goodman, A. L. Mapping human microbiome drug metabolism by
gut bacteria and their genes. Nature 570, 462–467 (2019).
61. Lillie, E. O. et al. The n-of-1 clinical trial: the ultimate strategy for
individualizing medicine? Per. Med. 8, 161–173 (2011).
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional ailiations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format,
as long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons license, and indicate
if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons license and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this license, visit http://creativecommons.
org/licenses/by/4.0/.
© The Author(s) 2023
Rosa Lundbye Allesøe1,2,3, Agnete Troen Lundgaard 1,2, Ricardo Hernández Medina 1, Alejandro Aguayo-Orozco1,2,
Joachim Johansen 1,2, Jakob Nybo Nissen1, Caroline Brorsson1,2, Gianluca Mazzoni1,2, Lili Niu 1,
Jorge Hernansanz Biel 1,2, Valentas Brasas1, Henry Webel1, Michael Eriksen Benros 3,4, Anders Gorm Pedersen 2,
Piotr Jaroslaw Chmura1,2, Ulrik Plesner Jacobsen 1,2, Andrea Mari5, Robert Koivula 6, Anubha Mahajan6,
Ana Vinuela 7,8, Juan Fernandez Tajes6, Sapna Sharma9,10,11, Mark Haid 12, Mun-Gwan Hong 13, Petra B. Musholt14,
Federico De Masi1,2, Josef Vogt15, Helle Krogh Pedersen2,15, Valborg Gudmundsdottir1,2, Angus Jones16, Gwen Kennedy 17,
Jimmy Bell18, E. Louise Thomas 18, Gary Frost 19, Henrik Thomsen20, Elizaveta Hansen20, Tue Haldor Hansen 15,
Henrik Vestergaard15, Mirthe Muilwijk21, Marieke T. Blom22, Leen M. ‘t Hart21,23,24, Francois Pattou25, Violeta Raverdy25,
Soren Brage26, Tarja Kokkola27, Alison Heggie28, Donna McEvoy29, Miranda Mourby30, Jane Kaye 30,
Andrew Hattersley 16, Timothy McDonald16, Martin Ridderstråle 31, Mark Walker32, Ian Forgie33,
Giuseppe N. Giordano34, Imre Pavo35, Hartmut Ruetten14, Oluf Pedersen 15, Torben Hansen 15, Emmanouil Dermitzakis7,
Paul W. Franks31,36,37, Jochen M. Schwenk 13, Jerzy Adamski38,39,40, Mark I. McCarthy6,41,42, Ewan Pearson33,
Karina Banasik1,2, Simon Rasmussen 1 , Søren Brunak 1,2 & IMI DIRECT Consortium
1Novo Nordisk Foundation Center for Protein Research, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark.
2Department of Health Technology, Technical University of Denmark, Kongens Lyngby, Denmark. 3Copenhagen Research Centre for Mental Health,
Mental Health Centre Copenhagen, Copenhagen University Hospital, Copenhagen, Denmark. 4Department of Immunology and Microbiology, Faculty of
Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark. 5C.N.R. Institute of Neuroscience, Padova, Italy. 6Wellcome Centre for
Human Genetics, University of Oxford, Oxford, UK. 7Department of Genetic Medicine and Development, University of Geneva Medical School, Geneva,
Switzerland. 8Biosciences Institute, Faculty of Medical Sciences, Newcastle University, Newcastle, UK. 9Research Unit of Molecular Epidemiology,
Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg, Bavaria, Germany. 10Institute of Epidemiology, Helmholtz
Zentrum München, German Research Center for Environmental Health, Neuherberg, Bavaria, Germany. 11Chair of Food Chemistry and Molecular and
Sensory Science, Technical University of Munich, Freising, Germany. 12Metabolomics and Proteomics Core, Helmholtz Zentrum Muenchen, German
Research Center for Environmental Health, Neuherberg, Germany. 13Afinity Proteomics, Science for Life Laboratory, School of Engineering Sciences in
Chemistry, Biotechnology and Health, KTH Royal Institute of Technology, Solna, Sweden. 14Research and Development Global Development, Translational
Medicine and Clinical Pharmacology, Sanoi-Aventis Deutschland, Frankfurt, Germany. 15Novo Nordisk Foundation Center for Basic Metabolic Research,
Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark. 16University of Exeter Medical School, Exeter, UK. 17The
Immunoassay Biomarker Core Laboratory, School of Medicine, University of Dundee, Dundee, UK. 18Research Centre for Optimal Health, Department
of Life Sciences, University of Westminster, London, UK. 19Section for Nutrition Research, Faculty of Medicine, Imperial College London, London, UK.
20Department of Radiology, Copenhagen University Hospital Herlev-Gentofte, Herlev, Denmark. 21Department of Epidemiology and Data Science,
Amsterdam Public Health Research Institute, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands. 22Department of General
Practice, Amsterdam Public Health Research Institute, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands. 23Department
of Biomedical Data Science, Section Molecular Epidemiology, Leiden University Medical Center, Leiden, the Netherlands. 24Department of Cell and
Chemical Biology, Leiden University Medical Center, Leiden, the Netherlands. 25Inserm, Univ Lille, CHU Lille, Lille Pasteur Institute, EGID, Lille, France.
26MRC Epidemiology Unit, University of Cambridge School of Clinical Medicine, Cambridge, UK. 27Department of Medicine, University of Eastern Finland,
Kuopio, Finland. 28Institute of Cellular Medicine, Newcastle University, Newcastle, UK. 29Diabetes Research Network, Royal Victoria Inirmary, Newcastle,
UK. 30Centre for Health, Law and Emerging Technologies (HeLEX), Faculty of Law, University of Oxford, Oxford, UK. 31Lund University Diabetes Centre,
Department of Clinical Sciences, Lund University, Malmö, Sweden. 32Translational and Clinical Research Institute, Faculty of Medical Sciences, Newcastle
University, Newcastle, UK. 33Division of Population Health & Genomics, School of Medicine, University of Dundee, Dundee, UK. 34Genetic and Molecular
Epidemiology Unit, Lund University Diabetes Centre, Department of Clinical Sciences, CRC, Lund University, SUS, Malmö, Sweden. 35Eli Lilly Regional
Operations, Vienna, Austria. 36Harvard T.H. Chan School of Public Health, Boston, MA, USA. 37OCDEM, Radcliffe Department of Medicine, University of
Oxford, Oxford, UK. 38Institute of Experimental Genetics, Helmholtz Zentrum München, German Research Center for Environmental Health, Neuherberg,
Germany. 39Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore. 40Institute of
Biochemistry, Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia. 41Oxford Centre for Diabetes, Endocrinology and Metabolism, University of
Oxford, Oxford, UK. 42Present address: Genentech, South San Francisco, CA, USA. e-mail: simon.rasmussen@cpr.ku.dk; soren.brunak@cpr.ku.dk
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
IMI DIRECT Consortium
Philippe Froguel25,43, Cecilia Engel Thomas1,2,13, Ragna Haussler13, Joline Beulens21, Femke Rutters21, Giel Nijpels21,
Sabine van Oort21, Lenka Groeneveld21, Petra Elders22, Toni Giorgino44,45, Marianne Rodriquez46, Rachel Nice47, Mandy Perry47,
Susanna Bianzano48, Ulrike Graefe-Mody49, Anita Hennige50, Rolf Grempler51, Patrick Baum51, Hans-Henrik Stærfeldt2,
Nisha Shah30, Harriet Teare30, Beate Ehrhardt52, Joachim Tillner53, Christiane Dings54, Thorsten Lehr54, Nina Scherer54,
Iryna Sihinevich54, Louise Cabrelli55, Heather Loftus55, Roberto Bizzotto5, Andrea Tura5, Koen Dekkers24, Nienke van Leeuwen24,
Leif Groop31, Roderick Slieker21,24, Anna Ramisch7, Christopher Jennison56, Ian McVittie29, Francesca Frau57,
Birgit Steckel-Hamann58, Kofi Adragni58, Melissa Thomas58, Naeimeh Atabaki Pasdar34, Hugo Fitipaldi34, Azra Kurbasic34,
Pascal Mutie34, Hugo Pomares-Millan34, Amelie Bonnefond25, Mickael Canouil25, Robert Caiazzo25, Helene Verkindt25,
Reinhard Holl59, Teemu Kuulasmaa60, Harshal Deshmukh28, Henna Cederberg61, Markku Laakso61, Jagadish Vangipurapu61,
Matilda Dale61, Barbara Thorand10,62, Claudia Nicolay63, Andreas Fritsche64, Anita Hill65, Michelle Hudson65, Claire Thorne65,
Kristine Allin15, Manimozhiyan Arumugam15, Anna Jonsson15, Line Engelbrechtsen15, Annemette Forman15, Avirup Dutta15,
Nadja Sondertoft15, Yong Fan15, Stephen Gough41, Neil Robertson41, Nicky McRobert41, Agata Wesolowska-Andersen41,
Andrew Brown33, David Davtian33, Adem Dawed33, Louise Donnelly33, Colin Palmer33, Margaret White33, Jorge Ferrer66,
Brandon Whitcher18, Anna Artati10, Cornelia Prehn10, Jonathan Adam10, Harald Grallert10,62, Ramneek Gupta2, Peter Wad Sackett2,
Birgitte Nilsson1,2, Konstantinos Tsirigos1,2, Rebeca Eriksen19, Bernd Jablonka67, Mathias Uhlen68, Johann Gassenhuber69,
Tania Baltauss70, Nathalie de Preville70, Maria Klintenberg34 & Moustafa Abdalla6
43Department of Metabolism, Digestion and Reproduction, Imperial College London, London, UK. 44Biophysics Institute (IBF-CNR), National Research
Council of Italy, Milan, Italy. 45Department of Biosciences, University of Milan, Milan, Italy. 46Biotech & Biomarkers Research Department, Institut de
Recherches Internationales Servier, Croissy sur Seine, France. 47Blood Sciences, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK. 48Boehringer
Ingelheim International, Therapeutic Area CardioMetabolism and Respiratory Medicine, Ingelheim am Rhein, Germany. 49Boehringer Ingelheim
International, Therapeutic Area CNS, Retinopathies and Emerging Areas, Ingelheim am Rhein, Germany. 50Boehringer Ingelheim International, Medicine
Cardiometabolism and Respiratory, Biberach an der Riss, Germany. 51Boehringer Ingelheim International, Translational Medicine & Clinical Pharmacology,
Biberach an der Riss, Germany. 52Centre for Mathematics and Algorithms for Data, University of Bath, Bath, UK. 53Clinical Operations, Sanoi-Aventis
Deutschland, Frankfurt, Germany. 54Clinical Pharmacy, Saarland University, Saarbrücken, Germany. 55Clinical Research Centre, Ninewells Hospital
and Medical School, University of Dundee, Dundee, Scotland, UK. 56Department of Mathematical Sciences, University of Bath, Bath, UK. 57Digital
and Data Sciences, Sanoi-Aventis Deutschland, Frankfurt, Germany. 58Eli Lilly and Company, Indianapolis, IN, USA. 59Institute for Epidemiology and
Medical Biometry, ZIBMT, University of Ulm, Ulm, Germany. 60Institute of Biomedicine, Bioinformatics Center, University of Eastern Finland, Kuopio,
Finland. 61Institute of Clinical Medicine, Internal Medicine, University of Eastern Finland, Kuopio, Finland. 62German Center for Diabetes Research,
München-Neuherberg, Germany. 63Lilly Deutschland, Bad Homburg, Germany. 64Medizinische Universitätsklinik Tübingen, Eberhard Karls Universität
Tübingen, Tübingen, Germany. 65NIHR Exeter Clinical Research Facility, University of Exeter Medical School, Exeter, UK. 66Regulatory Genomics and
Diabetes, Centre for Genomic Regulation, CIBERDEM, Barcelona, Spain. 67Strategy and Innovation, Sanoi-Aventis Deutschland, Frankfurt, Germany.
68Systems Biology, Science for Life Laboratory, School of Engineering Sciences in Chemistry, Biotechnology and Health, KTH Royal Institute of
Technology, Solna, Sweden. 69TMED, Sanoi-Aventis Deutschland, Frankfurt, Germany. 70Translational and Clinical Research, Metabolism Innovation Pole,
Institut de Recherches Internationales Servier, Suresnes Cedex, France.
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
Methods
The cohort
The cohort and available data included in the study are described
in detail in Koivula et al.
62,63
and Wesolowska–Andersen and Brors-
son et al. (ref. 7). In brief, we used the newly diagnosed sub-cohort
of the IMI-DIRECT study consisting of 789 participants. Fifty-eight
percent of participants was male and participants had the following
characteristics at baseline: age 62 (8.1) years; body mass index 30.5 (5.0)
kg m
−2
; fasting glucose 7.2 (1.4) mmol l
−1
; 2 h glucose 8.6 (2.8) mmol l
−1
.
Participants were diagnosed within 2 years before recruitment and
had glycated hemoglobin (HbA1c) < 60.0 mmol mol
−1
(<7.6%) within
the previous 3 months. All samples represent distinct individuals.
Furthermore, while Wesolowska–Andersen and Brorsson et al.
7
used
data from baseline and follow up at 18 and 36 months we only used
baseline data for modeling. In addition to the baseline data from Weso-
lowska–Andersen and Brorsson, we carried out extensive curation and
harmonization of the medication records included in the electronic
case forms by the research nurses in the different recruitment cent-
ers and thus used standardized ATC annotated medication data for
the individuals (see further detail below). Approval for the study pro-
tocol was obtained from each of the regional research ethics review
boards separately (Lund, Sweden: 20130312105459927; Copenhagen,
Denmark: H-1-2012-166 and H-1-2012-100; Amsterdam, Netherlands:
NL40099.029.12; Newcastle, Dundee, and Exeter, UK: 12/NE/0132)
and all participants provided written informed consent at enrollment.
The research conformed to the ethical principles for medical research
involving human participants outlined in the declaration of Helsinki.
Further details about the data generation can be found in Wesolowska–
Andersen and Brorsson et al.7.
Pre-processing of data
From the clinical, environmental, and questionnaire data only variables
with variation across the dataset that were present in at least 10% of
the individuals were included. The genomic data was included as the
genotypes of risk alleles identified in Mahajan et al.64. In total 393 risk
alleles were identified in our cohort out of the 403 associations men-
tioned in the paper. The genotypes were included as homozygous for
risk allele, heterozygote, not having the allele, or missing if the locus
was not identified for the individual. Diet data was included as 47
features on self-reported total intake of macronutrients and vitamins
across a 24-h period. The wearables measured with an accelerometer
included 25 measurements that summarize the movement and heart
rate during the day. Transcriptomics data (RNA sequencing) from
fasting whole blood samples were processed with RailRNA (v0.2.4b)
65
to obtain scaled counts for all samples and only the most variable
genes were included. The variable genes were selected by calculat-
ing the standard deviation across all individuals for each gene and
selecting genes with an above-average standard deviation. Both
targeted and untargeted metabolomics data in fasting plasma were
included for all measurements passing quality control. In the prot-
eomics data, all measurements within the measurable range based on
the OLINK antibody panel were included and residualized for plate
layout. The metagenomics data was only available for approximately
one-third (256) of the individuals and were included as normalized read
counts of identified Metagenomic Species
66
. Categorical data, includ-
ing questionnaire responses, drug data, and genomics, was one-hot
encoded. The continuous data were residualized by the collection
center as the data was collected from six different European countries
and, thus, handled by different nurses and lab technicians, as well as
differences in the time-of-day samples were taken, which could have a
large effect on the measurements. Additionally, the data were residual
-
ized for age and sex as these could be biological non-disease-related
confounders in the data. Lastly, each continuous dataset was z-scale
normalized per feature to ensure that each feature was distributed
around zero.
Classification of drugs using the ATC system
The ATC system is the WHO classification system for therapeutic
drugs. The system has a hierarchical structure, where the topmost
level, ‘level 1—Anatomical main group’, specifies the target organ or
tissue, and the lowermost level, ‘level 5—chemical substance’, specifies
the active chemical compound. The three levels in between specify
the therapeutic, pharmacological, and chemical levels, respectively.
We, therefore, mapped all drugs to the lowest possible level to prevent
information loss. A total of 4,155 entries could be mapped to level 5. For
55 entries, only a higher-level mapping was possible owing to lack of
specificity and 43 entries could not be mapped to the ATC system, either
because of the compound not existing in the database, for example
nutraceutical compounds, or when we were unable to identify which
drug was registered for the participant. The ATC system does not
only specify compound names, but also administration route and
daily dosages for over half of level 5 entries. However, owing to uncer-
tainty of the reliability of the registered dosages, only drug names
and administration routes were used for mapping. In instances where
the administration route was not available, the drug was mapped by
drug name only.
Drug data collection and clean-up
The study participants were asked to register their current drug usage
at screening and baseline. Drug names were registered as free text
together with administration route, dosage and frequency, and indica-
tion. Metformin was recorded separately from other anti-diabetic and
non-anti-diabetic drugs. The collected data was variable in quality,
using both generic and brand names, which were in many cases specific
to the country of the participant. The data was cleaned in four steps:
(1) removal of special characters, company names, formulations,
and other non-relevant information; (2) automatic mapping to the
PubChem database; (3) manual mapping to generic drug names; and
(4) mapping to the ATC system. Indications of placebo use, for example
participation in clinical drug trials, were noted as such. Only active
compounds were included and consequently, possible brand variation
was ignored, including for dietary supplements. Drug combinations
were mapped, when possible, to the ATC code specifying said combi-
nation. However, when the specificity of the proposed ATC code was
less specific than the registered drugs, the drug combinations were
mapped to individual ATC codes, that is, ‘Perindopril’ (C09AA04) and
‘Indapamide’ (C03BA11) was used instead of ‘Perindopril and diuretics’
(C09BA04). Entries were mapped to ATC codes with the administration
route when possible and otherwise mapped without the administration
route. Dosage information was not used in the mapping process. In the
manual mapping process, 99.4% of terms were assigned and a total of
359 drugs and drug combinations were identified. A total of 339 drugs
(94.4%) was mapped to 441 ATC codes.
Design of the VAE
The VAE framework was constructed to account for a variable number
of fully connected hidden layers in both the encoder and decoder and
a latent layer that samples from a Gaussian distribution N(0, 1) of two
vectors of size N
L
representing the means, µ, and standard deviations, σ.
Each hidden layer included both batch normalization and dropout
67
and with leaky rectified linear units (LeakyReLU)68 as activation func-
tion. Each dataset was concatenated to one input layer of both categori-
cal and continuous variables. To allow for dataset-specific weights the
error calculation was done separately for each dataset. Here we applied
cross-entropy loss for categorical data and mean squared error for
continuous data as implemented in PyTorch69. The loss was normal-
ized by dataset input size and batch size. Deviance from the Gaussian
distribution was penalized by adding the Kullback–Leibler divergence
(KLD) to the loss. The final loss was defined as
L=Wcat ×Ecat +Wcon ×Econ +WKLD ×KLD
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
Here, Ecat and Econ are vectors of normalized reconstruction
error for each of the continuous and categorical datasets. W
cat
and
W
con
are vectors as well of the same length as the errors to introduce
dataset-specific weights. We applied an equal weight of 1 for all datasets
except for continuous clinical data where we used a weight of 2. WKLD
is a weight put on the KLD defined as WKLD = β × NL−1 for which we used
a β of 0.0001 for the final model. The KLD was defined as
KLD =∑−
1
2
(1+ln (σ)−μ2−σ)
To efficiently handle missing data for the continuous features we
encoded them as mean values across a particular feature during training
and excluded the missing data points during back-propagation. With
the data being z-score normalized the mean value is represented as zero.
For the categorical features, we included them as a zero vector and the
ignore index feature in the cross-entropy implementation in PyTorch
was used to not include errors for missing data in the back-propagation.
The VAE model was trained with the Adam optimizer70, with a mini-batch
size of 10 and increasing batch size with a factor of 1.25 during training
after every 50 epochs. The number of training epochs was set to 200 on
the basis of early stopping on the test set as described below. Additionally,
we trained the model using warm-up by first including the full KLD after
10 epochs slowly increasing the weight at epochs 4, 6, and 8. The latent
representation of each patient was obtained by passing them through
the trained VAE and extracting the µ layer. The VAE was implemented
using PyTorch
69
(v.1.7.0) and run using a GPU running CUDA (v.10.2.89).
Hyperparameter optimization for multi-omics integration
We initially divided the dataset into training (90%) and test (10%) sets to
identify the optimal hyperparameter settings to efficiently capture the
data structure without losing the ability to generalize on the test data
(Supplementary Figs. 2 and 3). We tested different combinations of
sizes of hidden layers, the number of hidden layers, size of latent space,
dropout, and weight on the KLD. We then evaluated the model on the
basis of both test log-likelihood and reconstruction accuracy. For the
number of hidden neurons, the variations used were 200, 500, 800,
1,000, and 1,200, with the number of layers ranging between 1 and 5. The
tested latent sizes were between 20 and 400 as well as dropout of 10%,
20%, and 30% and KLD weights of 0.001, 0.0001, and 0.0001. We defined
an accurate reconstruction for categorical variables as the class with
the highest probability corresponding to the class given by the input.
For continuous variables, the accuracy was assessed by comparing the
reconstructed array with the input array using cosine similarity for each
individual instead of using exact matching. For both categorical and
continuous data only non-missing values were used when calculating
the accuracy in the reconstruction. We chose the number of training
epochs on the basis of when the optimal test likelihood was achieved
during testing rounded up to the nearest 100 epochs to ensure sufficient
training to learn the complexity of the data. Here we found that more
complex models, with higher numbers of hidden neurons and layers,
resulted in worse performance on the test set (Supplementary Fig. 2) and
that models with more than one hidden layer were unable to provide a
decent reconstruction on the training data without overfitting. The only
exception was the size of the latent representation, which gave a worse
performance with smaller sizes (<50) and equally good performance for
larger sizes (from 100 to 400) (Supplementary Fig. 3). For the five best
performing models, stability was measured to choose the final model. The
stability of the model was evaluated by repeating training with the same
hyperparameters and calculating the difference in cosine similarity of
the latent space to all other individuals. If the model produced the same
result the average change in cosine similarity should be zero. The model
with the average change closest to zero was then considered the most
stable. The final hyperparameters were set to be one hidden layer of 2,000
neurons, a latent size of at least 100, and a 10% dropout for regularization.
Evaluating feature importance
Feature importance was extracted from the weights of the network for
the models with only one hidden layer and because the input data was
z-score normalized calculated as
Ii
n
hidden
j1
wij
where I
i
is the ith feature input and
|
|wij|
|
is the absolute value of the weight
from ith input to the jth hidden neuron. To assess the actual impact on
the latent representation an adaptation of the SHAP19 analysis was
applied. The difference in model performance was assessed as the abso-
lute differences of the latent representation when changing each input
to missing for all individuals and passing it through the trained model.
Extracting significant drug associations
Drug associations were extracted by perturbation of the input data
after training the final model on all individuals. Thus, for each drug we
changed the drug status for all individuals with ‘not receiving’ to ‘receiv-
ing’. Importantly, we only included individuals that did not receive the
specific drug or another drug within the same therapeutic subgroup
(ATC level 2). Then, for each drug change, we compared the change in
reconstructions to when we passed the original (un-perturbed) data
through the network. In other words we determined the differences
that the network infers from the change in drug status that during train-
ing was learned from all individuals receiving the drug. We used two
strategies for this, one was based on an ensemble of Student’s t-tests
using benchmarked thresholds, and another was based on Bayesian
decision theory. Both approaches were benchmarked against rand-
omized datasets where all the input data matrices were shuffled on rows
and columns. We simulated effects in the shuffled data by randomly
sampling a combination of a drug, a multi-omics dataset, and a feature
within that omics dataset. For each combination, we then sampled
an effect from the standard normal distribution N(0,1) and added
this value to the omics feature whenever the selected drug was taken
by an individual. We, therefore, did not expect that all effects would
be significant in the statistical tests because we sample from N(0,1)
and some effects will be close to 0. We added a total of 100 effects to
the shuffled data and repeated the entire procedure to generate two
shuffled datasets each with their unique added effects. Additionally,
we investigated if the number of significant associations, effect size
estimates and model uncertainty in the reconstruction were not biased
by individual dataset uncertainties. This was done by calculating PCCs
between the average estimated effect size across all 20 drugs and the
difference between model input and the reconstructions for each of
the omics features.
Significant associations using MOVE t-test
To evaluate if the change in the reconstruction was significant, we first
determined the expected average change when passing the original
and perturbed data through the model ten times. On the basis of these
averages, we used a Student’s t-test for related samples as implemented
in Python SciPy (v.1.3.1)71 between the baseline and drug-perturbed data
for all non-missing continuous data. All P values were subsequently
Bonferroni-corrected independently for each drug, and we applied
a significance threshold of adjusted P < 0.05. We repeated the entire
analysis with retraining of the model 10 times for each of four latent
sizes (150, 200, 250, and 300). Associations were only included for
analysis if they were significant for at least three of the four latent
sizes and in at least five out of ten of the repeats. Therefore, reported
P values were the averaged P value across the 10 replicate and 4 model
tests, that is a total of 40 two-sided Bonferroni-corrected t-tests. The
change in reconstruction, what we report as effect size, was calculated
as the average difference across the 10 replicates and 4 model tests and
were reported with 95% confidence intervals.
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
Significant associations using Bayes decision theory
For the method that was based on Bayesian decision theory we used
an approach inspired by single-cell variational inference29 and Lopez
et al.
31
. We trained VAE models with a latent size of 150 neurons and
benchmarked the approach using different latent sizes and ensembling
1, 5, 10, 20, 30, 35, 40, or 50 models, which we termed refits. For the
refits we averaged the reconstructions and used these to obtain the
posteriors for the non-perturbed data and each of the drug perturba-
tions. Thus, for VAE ensemble refit i, individual n, feature f, and drug d
we define the variational reconstructions as
xinfd
. By averaging across
VAE refits, we obtain estimates of the average posteriors
xnfd
. Then, for
each drug d we compare between two models:
Mf
d
where feature f is
significantly associated with the drug, and the alternative model
Mf
0
where feature f is not significantly associated with drug d. Hence, we
evaluate how often
|
|
xnfd −
xnf0|
|>0
and calculate Bayes factors (K) as:
Kloge
PM
f
d
xfd
xf0
PM
f
0
xfd
xf0
We ranked the associated features according to K (ref. 72). We set
a FDR of α by accepting associations (n) between features and a drug
until the cumulative evidence of P(M
0
) across accepted features for
the drug was above the threshold. Since
P(Mf
0)=(1−P(Mf
d))
we accepted
drug-feature associations while the cumulative evidence E is lower
than α
E=∑
f
(1−P(M
f
d))
n<α
Benchmarking of t-test, MOVE t-test and MOVE Bayes
To be able to compare the number of significant associations between
methods we used the two randomized datasets to estimate FDR from
the ground truth, that is the added drug–omics effects (Supplementary
Table 3). Here we found that a t-test with Benjamini–Hochberg FDR of
0.01 had ground-truth FDR of 0.00 and 0.06 on the two randomized
datasets, corresponding to 52 and 67 true positives as well as 0 and 4
false positives, respectively. For MOVE t-test, we benchmarked the num-
ber of refits of the 4 models and found 10 refits to have a ground-truth
FDR of 0.02 and 0.06, with 48 and 61 true positives as well as 1 and 3 false
positives, respectively. For MOVE Bayes we benchmarked the number
of refits for a model with 150 latent neurons and found FDR from the
cumulative evidence to be well aligned with FDR of the ground truth.
Using Bayes FDR of 0.05 we found 30 refits to have ground-truth FDR of
0.02 and 0.05, respectively. Across the two shuffled datasets 42 and 59
true positives were found by all three methods (Supplementary Fig. 12).
Calculation of drug associations using other methods
We compared our findings to associations identified with standard
statistical approaches using Student’s t-test for unrelated samples and
an ANOVA between two groups of individuals ‘not receiving’ and ‘receiv-
ing’ each drug. Here we used Benjamini–Hochberg correction for FDR
73
with an adjusted P < 0.01. Additionally, we tested if a least absolute
shrinkage and selection operator (LASSO) model was able to identify
features with significant impact on predicting the ‘not receiving’ or
‘receiving’ groups for each drug. However, the LASSO model was unable
to converge possibly owing to the high input feature dimensionality.
All statistical tests were done with Python SciPy (v.1.3.1)71.
Drug effect size and similarities across omics data
Drug effect sizes were determined as the difference between the base-
line and drug-perturbed variational reconstructions, that is, as the
average difference across the VAE ensemble refits reported with 95%
confidence intervals. Drug similarities were calculated as the cosine
similarity as implemented in Python SciPy (v.1.3.1)
71
between the aver-
age effect sizes on all features identified as significantly associated
for at least one of the drugs both across and within each dataset. The
difference was only calculated for non-missing data and individuals not
already on the drug or a drug in the same ACT group. The rank of drug
effect sizes was determined for each omics dataset ranking the effect
sizes from 1 to 20. A rank of 20 indicates that the drug had the highest
average effect size in this omics dataset compared to the other drugs.
Correlations between multi-omics profiles and number of individuals
taking the drug pair were calculated from the fraction of individuals
that overlapped between the two drugs.
Molecular-focused analysis of the multi-omics data
To get a better understanding of the molecular profiles identified in the
associations for the transcriptomics and proteomics data we tested for
enriched Gene Ontology terms as well as molecular pathways. For the
transcriptomics data, we assessed the molecular patterns of biological
processes and pathways from Reactome
74
(v.3.7) using the significantly
associated genes for each drug against a background list of all genes
included in the data integration. We used WebGestaltR
75
(v.0.4.4) for
the analysis with default settings (hypergeometric test) and evaluated
all results with an FDR < 0.05. The targeted metabolomics data was
analyzed for potential metabolite enrichments using MetaboAnalyst
76
(v.5) over-representation analysis using a hypergeometric test and
FDR of 0.05. We investigated both enrichments in known pathways in
the KEGG database as well as enrichment of chemical structures sub-,
main- and super-class levels. For all analyses, we used the included
panel of targeted metabolites as the reference data.
Association differences within diabetes archetypes
As mentioned, previous work by Wesolowska–Andersen and Brorsson
et al. performed archetype analysis of the multi-omics data with only
metformin medication data7. Here they based the archetypes on clini-
cal markers and identified four distinct and one ‘mixed’ T2D arche-
types with clinical and omics profiles. To investigate if these distinct ar
chetypes differed in their drug associations we used a t-test on the
average effect size change for the individuals of each archetype against
the remaining individuals. The analysis was only done for the significant
drug associations for each drug. All analysis was only done for individu-
als not taking the drug or a drug within the same ATC therapeutical class
similarly to the main analysis.
Drug–drug interactions
We used an in-house drug–drug interaction compendium generated
from publicly available sources (Supplementary Table 11) to assess
whether drug combinations had been reported previously to be
interacting or not77. The compendium contains interactions from
26 different datasets of pharmacovigilance, clinically oriented infor-
mation, schemas for NLP corpora, and drug–Cytochrome P450 rela-
tionships sources. For 12 of the drug–drug pairs in our dataset we
could identify drug–drug interactions with reported severity (major,
moderate, minor, possible, undetermined, and none) indicating
clinical significance.
Reporting summary
Further information on research design is available in the Nature
Research Reporting Summary linked to this article.
Data availability
Owing to the informed consent given by study participants, the vari-
ous national ethical approvals for the present study, and the Euro-
pean General Data Protection Regulation (GDPR), individual-level
clinical and omics data cannot be transferred from the centralized
IMI-DIRECT repository. Requests for access to summary statistics of
the IMI-DIRECT data, including those presented here, can be made to
Nature Biotechnology
Article https://doi.org/10.1038/s41587-022-01520-x
directdataaccess@dundee.ac.uk. Requesters will be informed on how
summary-level data can be accessed via the DIRECT secure analysis
platform following submission of an appropriate application. The
IMI-DIRECT data access policy is available at https://directdiabetes.org.
Example data is available at https://github.com/RasmussenLab/MOVE/
for testing of MOVE. As described in the methods section we used
ATC (https://www.who.int/tools/atc-ddd-toolkit/atc-classification)
and WebGestalt (v.0.4.4 at http://www.webgestalt.org) for analysis of
Gene Ontologies, Reactome (v.3.7 at https://reactome.org) for analy-
sis of molecular pathways, and MetaboAnalyst (v.5 at https://www.
metaboanalyst.ca) for analysis of targeted metabolomics data. The
25 databases of drug–drug interactions are listed in Supplementary
Table 11. Source data are provided with this paper.
Code availability
The MOVE pipeline is available at https://github.com/RasmussenLab/
MOVE/.
References
62. Koivula, R. W. et al. Discovery of biomarkers for glycaemic
deterioration before and after the onset of type 2 diabetes:
rationale and design of the epidemiological studies within the IMI
DIRECT Consortium. Diabetologia 57, 1132–1142 (2014).
63. Koivula, R. W. et al. Discovery of biomarkers for glycaemic
deterioration before and after the onset of type 2 diabetes:
descriptive characteristics of the epidemiological studies within
the IMI DIRECT Consortium. Diabetologia 62, 1601–1615 (2019).
64. Mahajan, A. et al. Fine-mapping type 2 diabetes loci to
single-variant resolution using high-density imputation and
islet-speciic epigenome maps. Nat. Genet. 50, 1505–1513 (2018).
65. Nellore, A. et al. Rail-RNA: scalable analysis of RNA-seq splicing
and coverage. Bioinformatics 33, 4033–4040 (2017).
66. Nielsen, H. B. et al. Identiication and assembly of genomes and
genetic elements in complex metagenomic samples without
using reference genomes. Nat. Biotechnol. 32, 822–828 (2014).
67. Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. &
Salakhutdinov, R. Dropout: a simple way to prevent neural net-
works from overitting. J. Mach. Learn. Res. 15, 1929–1958 (2014).
68. Maas, A. L., Hannun, A. Y. & Ng, A. Y. Rectiier nonlinearities
improve neural network acoustic models. In Proc. 30th
International Conference on Machine Learning (eds Dasgupta, S. &
McAllester, D.) (JMLR, 2013).
69. Paszke, A. et al. PyTorch: an imperative style, high-performance
deep learning library. In Advances in Neural Information Processing
Systems Vol. 32 (eds Wallach, H. et al.) (Curran Associates, 2019).
70. Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization.
Preprint at arXiv https://doi.org/10.48550/arXiv.1412.6980 (2014).
71. Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientiic
computing in Python. Nat. Methods 17, 261–272 (2020).
72. Kass, R. E. & Raftery, A. E. Bayes factors. J. Am. Stat. Assoc. 90,
773–795 (1995).
73. Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate:
a practical and powerful approach to multiple testing. J. R. Stat.
Soc. Ser. B Stat. Methodol. 57, 289–300 (1995).
74. Jassal, B. et al. The reactome pathway knowledgebase. Nucleic
Acids Res. 48, D498–D503 (2020).
75. Liao, Y., Wang, J., Jaehnig, E. J., Shi, Z. & Zhang, B. WebGestalt
2019: gene set analysis toolkit with revamped UIs and APIs.
Nucleic Acids Res. 47, W199–W205 (2019).
76. Chong, J. & Xia, J. MetaboAnalystR: an R package for lexible and
reproducible analysis of metabolomics data. Bioinformatics 34,
4313–4314 (2018).
77. Leal Rodríguez, C. et al. Drug interactions in hospital prescriptions
in Denmark: prevalence and associations with adverse outcomes.
Pharmacoepidemiol. Drug Saf. 31, 632–642 (2022).
Acknowledgements
We are grateful to IMI-DIRECT study participants who volunteered
for phenotyping, and clinical and technical sta across involved
European study centers who contributed to recruitment and clinical
assessment of study participants. The work leading to this publication
has received support from the Innovative Medicines Initiative Joint
Undertaking under grant agreement 115317 (DIRECT), resources of
which are composed of inancial contribution from the European
Union’s Seventh Framework Programme (FP7/2007-2013) and EFPIA
companies’ in kind contribution. Information on the initiatives and
activities of the IMI-DIRECT research consortium is available at
https://directdiabetes.org. R.L.A., A.T.L, R.H.M., J.J., V.B., H.W., J.N.N.,
C.B., G.M., L.N., P.J.C., U.P.J, K.B., S.R., and S.B. were supported
also by the Novo Nordisk Foundation (grants NNF14CC0001 and
NNF17OC0027594). Additionally, R.H.M. was supported by Novo
Nordisk Foundation (grant NNF20SA0035590). Figure 1a was partly
created using BioRender.com. Finally, we would like to thank Tugce
Karaderi for critical comments on the manuscript.
Author contributions
S.R. and S. Brunak designed and supervised the analyses, interpreted
the results, and wrote the manuscript. R.L.A. wrote the code, carried
out the analyses, interpreted the results, and wrote the manuscript.
A.T.L. cleaned and prepared the drug data to be used in the analysis.
R.H.M. developed the Bayesian analysis. J.J., J.N.N., V.B. and H.W.
assisted in the development of the method and analysis. A.A-O. assisted
in interpretation of the drug interactions and writing the manuscript.
G.M. processed transcriptomics data and L.N. assisted in interpreting
the proteomics results, C.B., K.B., M.E.B. and A.G.P. assisted in the
interpretation of results and design of the analysis. J.H.B. analyzed
drug–drug interaction data. Data acquisition and pre-processing was
performed by A.J., A. Mari, A.T., A. Mahajan, A.V., E.H., H.K.P., H.T., J.F.,
J.V., M.H., M.-G.H., P.B.M., G.K., J.B., L.T., G.F., R.K., S.S., T.M., T.K., and V.G.
Clinical investigations was performed by A. Hattersley, A. Heggie, D.M.,
F.P., F.R., G.N., P.E., T.H.H., H.V., T.H., H.T., M.R., and V.R. High-performance
computing support and administration was performed by P.J.C. and
U.P.J. Project administration was performed by F.D.M., I.F., J.K., R.K.,
G.N.G., I.P., H.R., O.P., M.M., M.W., E.P., S. Brage, and P.W.F. Funding
acquisition was by M.W., O.P., E.D., P.W.F., J.M.S., J.A., E.P., M.I.C., and
S. Brunak. All authors reviewed and edited the inal manuscript.
Competing interests
S. Brunak has ownerships in Intomics A/S, Hoba Therapeutics Aps,
Novo Nordisk A/S, Lundbeck A/S, and managing board memberships
in Proscion A/S and Intomics A/S. M.I.C. has served on advisory panels
for Pizer, Novo Nordisk, and Zoe Global; has received honoraria from
Merck, Pizer, Novo Nordisk, and Eli Lilly; and has received research
funding from Abbvie, Astra Zeneca, Boehringer Ingelheim, Eli Lilly,
Janssen, Merck, Novo Nordisk, Pizer, Roche, Sanoi Aventis, Servier,
and Takeda. As of June 2019, M.I.C. is an employee of Genentech and
a holder of Roche stock. E.P. has received honoraria from Sanoi and
Lilly. The other authors declare no competing interests.
Additional information
Supplementary information The online version contains supplemen-
tary material available at https://doi.org/10.1038/s41587-022-01520-x.
Correspondence and requests for materials should be addressed to
Simon Rasmussen or Søren Brunak.
Peer review information Nature Biotechnology thanks Yasuhiro Kojima
and Elin Nyman for their contribution to the peer review of this work.
Reprints and permissions information is available at
www.nature.com/reprints.
