Muhammad Arif's PhD Defense https://muharif.net 11 June 2021 KTH Royal Institute of Technology, Sweden Science For Life Laboratory

1. Systems and Network-based
Approaches to Complex
Metabolic Diseases
Muhammad Arif
Science for Life Laboratory, KTH Royal Institute of Technology
Supervisors: Prof. Dr. Adil Mardinoglu; Prof. Dr. Mathias Uhlén
Stockholm, 11 June 2021

2. • Need energy to be able to
perform activities
• Chemical à Kinetic
• Complex System
• Interconnected
Human Body == Car
2

3. Metabolites
Metabolomics
DNA
Genomics
RNA
Transcriptomics
Proteins
Proteomics
Microbiome
Metagenomics
3

4. Metabolites
Metabolomics
DNA
Genomics
RNA
Transcriptomics
Proteins
Proteomics
Microbiome
Metagenomics
4
Systems Biology
Paradigm
(Definition taken from Institute of Systems Biology)
• Holistic approach to answer
complex and important biological
questions
• Collaborative effort from
multiple disciplines
• Biology
• Computer Science
• Physics, etc
• Predictive analysis to understand
the condition changes

5. Approaches in Systems Biology
Statistical Inference
Network Analysis
Machine Learning
Omics
Data
Altered analytes
Functional Analysis
Classification
Regression
Clustering
Relationships
Centrality
Community
Patient characterization
Disease mechanism
Novel biomarkers
Novel Therapy
Drug Repositioning
5

6. Present Investigation
I II
Generation of
Biological Networks
III
Systems Biology of
Heart
IV
Systems Biology of
Muscle
V VI
Systems Biology of
Liver
6

7. Paper II
iNetModels 2.0: an interactive visualization
and database of multi-omics data
Arif and Zhang et. al. (2021)
Nucleic Acid Research
doi: 10.1093/nar/gkab254
7

8. Study Introduction
• More and more personalized multi-omics data were collected
• Integration of multi-omics data has been proven to offer novel
insights and comprehensive understanding of human body
• Problem: Limited studies in collecting and exhibiting data
association in a single database
• We generated integrated multi-omics networks from multiple
studies and conditions
• Goal: A database and interactive platform to visualize multi-
omics data interactions
8

9. Platform Description
Tissue-specific (GTEx)
Cancer-specific (TCGA)
Personalized Multi-Omics
Profiling
(6 sources)
Data Sources
Co-Expression Network
(Spearman Correlation)
Low Expression Filters
Age and Sex Correction
Network Generation
Database and Visualization
Cross and Delta Networks
Tissue; Cancer; Sex; Diseases
Statistical & Omics Filtering
Integration with other tools
Programmatic Access
Features
https://inetmodels.com
9

10. Use Case: NAFLD CMA Supplementation
Hypothesis Testing
Relationship between the
supplement with TG and liver
enzymes
Exploratory Analysis
Relationship between the
supplement with gut microbiomes
Results Validation
The effect of the supplement to
BCAA metabolism and glucose level
New Insights
CMA supplementation affects
several cholesterol-related variables
and inflammation markers
Source:
P100 Study
SCAPIS-SciLifeLab networks
10

11. Summary
Personalized Wellness
Profiling Studies
Multi-Omics iNetModels 2.0
The Cancer Genome
Atlas (TCGA)
Genotype-Tissue
Expression (GTEx)
Data Sources
https://inetmodels.com
11
• >100 Networks
• Flexible Customization

12. Paper III
Integrative transcriptomic analysis of tissue-
specific metabolic crosstalk after myocardial
infarction
Arif and Klevstig et. al. (2021)
eLife
doi: 10.7554/eLife.66921
12

13. Study Introduction
• Multiple studies have been performed and provided new
insights into MI
• Limitation: Single Tissue analysis
• Cross-talk between different tissues and their dysregulation has
not been examined
• In this study, we performed integrated analysis between heart
and metabolically active tissues
• Goal: More complete picture of metabolic alteration during MI
13

14. Study Flow
14

15. Time Series Analysis
Gene Ontology Reporter Metabolites
15

16. Co-expression Network Analysis
16

17. Co-expression Network Analysis
Autophagy
Endocytosys
(FoxO, Inslin, mTOR, AMPK) Signalling
Cell Cycle
Circadian Rhythm
Fatty acid metabolism
Amino Acid Transport
TCA Cycle
Tight Junction
m/RNA metabolism
Endosomal Transport
(Wnt, NFK-Beta) Signaling
(Retinol, Cholesterol,-
Fructose and Mannose,-
Fatty Acid, Steroid) Metabolism
Heart-Specific Functions
Oxytocin signalling
(Glycogen, Inositol phosphate,-
Purnine) Metabolism
Central Clusters
17

18. Final Results
Fatty Acid
Fatty Acid
Retinol
Lipid Metabolism (Up)
Inflamatory Response (Up)
Fatty Acid Metabolism (Down)
Lipid Metabolism (Up)
Inflamatory Response (Up)
Fatty Acid Beta-Oxidation (Up)
Glutathione Metabolism (Down)
Inflamatory Response (Down)
Fatty Acid Metabolism (Down)
Response to Lipid (Up)
Inflamatory Response (Up)
Retinoid metabolic process (Up)
Mitochondrial
Dysfunction
• We identified several
targets/biomarkers:
• Flnc
• Lgals3
• Prkaca
• Pprc1
Hypothesized Metabolic Cross-talk
18

19. Paper V
Multi-omics analysis reveals the influence of the
oral and gut microbiome on host metabolism in
non-alcoholic fatty liver disease
Zeybel and Arif et. al. (2021)
Manuscript
19

20. Study Introduction
• NAFLD has been labelled as “the silent
pandemic”
• One of the most prevalent diseases in the world (25%
of population)
• No approved treatment for this disease
• Dysbiosis of microbiomes have been suspected
to influence NAFLD
• Goal: systematic analysis to study the dysbiosis
of microbiomes and their relationships with
other omics
20

21. Study Design
No steatosis Mild steatosis Moderate steatosis Severe steatosis
Measure
Group
HS< 5.5% 5.5%≤HS<8% 8%≤HS<16.5% HS≥16.5%
MRI-PDFF
n=10 n=14 n=20 n=12
Blood
Feces
Saliva
21

22. Multi-Omics Data Integration
The network was retrieved from
iNetModels
22

23. Multi-Omics Data Integration
• Glutathione-related metabolites
associated with GGT
23

24. Multi-Omics Data Integration
• Glutathione-related metabolites
associated with GGT
• Known NAFLD-marker proteins
were positively correlated with
liver fat and enzymes
24

25. Multi-Omics Data Integration
• Glutathione-related metabolites
associated with GGT
• Known NAFLD-marker proteins
were positively correlated with
liver fat and enzymes
• Negative correlation of
important microbes to liver fat
25

26. Multi-Omics Data Integration
• Glutathione-related metabolites
associated with GGT
• Known NAFLD-marker proteins
were positively correlated with
liver fat and enzymes
• Negative correlation of
important microbes to liver fat
• Protagonist and NAFLD-
associated gut microbes
associated to ALT, AST, and uric
acid
26

27. Summary
• Multi-omics data from well-characterized NAFLD patients with
different hepatosteatosis severity levels
• Implementation of a wide range of systems biology approaches
• Single-omics analysis: Finding molecular signatures from each omics
type
• Multi-omics integration: functional relationships between analytes
from different omics types
• Elucidating the dysbiosis of microbiomes caused by NAFLD
• Identification of candidate novel biomarkers for NAFLD
27

28. Summary and Concluding Remarks
• Systems biology is a great tool to get a holistic and systematic
view of human body
• One of the main enabler and driver of personalized medicine
• Development and application of systems biology tools in
complex diseases using multi-tissue and multi-omics
data
28

29. Future Perspectives
• More personalized multi-omics studies
• Account for individual variation in healthy and disease state
• Lead towards better patient characterizations and biomarkers discovery
• Incorporation of prior knowledge to the networks
• To be able to derive causality from the network
• To shorten the analysis cycle
• General (and open) framework for data collection and analysis
• More robust disease model à Data, Data, and Data!
29

30. Open
Science
Open
Data
Open
Access
Open
Source
30
Adapted from:
DOI: 10.3233/ISU-170846

31. Acknowledgements
Adil Mardinoglu
Cheng Zhang
Woonghee Kim
Ozlem Altay
Xiangyu Li
Mengnan Shi
Hong Yang
Meng Yuan
London:
Stephen Doran
Simon Lam
Abdulahad B.
Ali Kaynar
Ex-Members:
Sunjae Lee
Rui Benfeitas
Alen Lovric
Natasa Sikanic
Dorines Rosario
Beste Turanli
Mohammed A.
Feride Eren
Mathias Uhlén
Linn Fagerberg
Max Karlsson
Abdelah Tebani
Wen Zhong
Jan Borén
Martina Klevstig
Malin Levin
Elias Björnson
Bash Biotech
Saeed Shoaie
And many others!
31
Jens Nielsen

32. 32