LUND
UNI
VERSI
TY
PO
Box
117
221
00
Lund
+46
46-
222
00
00
Enhancing prediction and causal inference in metabolic dyshomeostasis
Atabaki Pasdar, Naeimeh
2020
Document Version:
Publisher's PDF, also known as Version of record
Link to publication
Citation for published version (APA):
Atabaki Pasdar, N. (2020). Enhancing prediction and causal inference in metabolic dyshomeostasis. Lund
University, Faculty of Medicine.
Total number of authors:
1
General rights
Unless other specific re-use rights are stated the following general rights apply:
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors
and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the
legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study
or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal
Read more about Creative commons licenses: https://creativecommons.org/licenses/
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove
access to the work immediately and investigate your claim.
Enhancing prediction and casual inference
in metabolic dyshomeostasis
NAEIMEH ATABAKI-PASDAR
Department of Clinical Research, Malmö
Lund University, Faculty of Medicine
Doctoral Dissertation Series 2020:138
ISBN 978-91-8021-005-8
ISSN 1652-8220
9
789180
210058
Naeimeh Atabaki-Pasdar completed her
BSc in Biomedical Engineering at Amirkabir
University of Technology and her MSc in
Bioinformatics at Lund University. Naeimeh
has completed her PhD at the Genetic and
Molecular Epidemiology unit at the Lund
University Diabetes Centre. Naeimeh´s
doctoral thesis focuses on enhancing
pre-diction and causal inference in metabolic
dyshomeostasis.
≥
↔
−
−
−
−
−
−
−
≥
/
<
/
≥
≥
/
≥
−
!
=
(
− ˆ )
λ
β
!
=
(
−ˆ ) +λ×
!
=
|β |
=
+
+
+
+
=
+
=
+
=
+
=
×
×
+
[(
− [ ])( − [ ])]
=
"
χ
−
"
× ( − )
=
χ
(
)
−
χ
(
)
χ
(
)
−
=
(
)
− (
)
(
)
,
χ
−
(
∩ )
( )
B
A
D
C
E
( , , , , ) =
(
| , ) × ( | ) × ( | ) × ( ) × ( )
Θ =
{θ , ..., θ }
( )
(
| , ) × ( | ) ×
(
| ) × ( ) × ( )
(
|Θ) =
#
=
(
| ( ), θ )
χ
U
U
U
O
U
E
U
IV
☓
☓
−
−
−
−
β
−
=
β
−
β
−
±
=
|β
−
|
$
(
−
β
−
) + (
−
β
−
)
β =
!
=
β
ˆ
!
=
,
ˆ
β
α
β
α
β
α
β
α
β
− β
α
α
α
β
−β
γ
(γ)
−
α−
α
α
−β
≈
γ
(γ)
−
−
α≤
≤
χ
χ
χ
→
↑
→
↓
→
↓
→
↓
→
↑ →
↓
→
↓ →
↓ →
↑ →
↓
→
↑ →
↓ →
↑
→
↓ →
↑
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 10 20 30 40 Non−diabetic Diabetic Diabetes status Liv er f at percentage Non-diabetes Diabetes ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 10 20 30 40
Amsterdam (NL) Copenhagen (DK) Dundee (UK)
Exeter (UK) Kuopio (FI) Newcastle (UK) Center Liv er f at percentage
Cohort
B
A
Center
Non-diabetes
Diabetes
Li
ver
fa
t (%
)
Amste rdam( NL) Cope nhag en(D K) Dund ee(UK ) Exet er(U K) Kuop io(FI) New castl e(UK)×
−
T-M
116
E-P
377
T
16,209
G
108
C
58
Data reduction
(LASSO/Pairwise correlation)
T-M
25
E-P
22
T
93
G
23
C
22
Feature selection data
n = 735
Fatty liver (% yes) = 34
Data reduction
(LASSO)
U-M
39
T-P
48
Feature selection data
Complete case
T-P data n=364
U-M
172
483
T-P
Complete case
U-M data n=377
0.67 0.7 0.72 0.74 0.79 0.8 0.8 0.81 0.82 0.82 0.84 0.5 0.6 0.7 0.8 0.9 1.0 Genetic(G) Metabolomic(M)Transcr iptomic(T) Proteomic(P) Clinical(C) M+C P+C T+C G+T+M+P G+C C+G+T+M+P Models R OCA UCM
P
M
P
r132 r448 PC..C44.4r614 PC..C40.0lPC..C18 r222 PC..C44.5FCRL ACTRr563 GAB1 RSPO PC..C42.5GNAO TRIM AC09 PC.ae.C40.3ZSWI RAB3DSP DNAHUCP3 FBXO39CPT1 PC.aa.C40.3XIAP C10 FADS COBLr121 SDHA ELMO PPP1R10UBE2 PGF_ PC..C42.0YY1A CRB2r740 r116 AFAPTTLL FOLR CLECIFI2 TMEM2 CCDC11SFT2 AC003NFKB CCR2 ARHG H40_ PC..C34.3MIR6 ENSG0000020TBC1 PMP2 ENSG0000023PRSS SBP TMEM7 LOC100r354 CNIH NOC4 PC..C34.2 LOC105372CIB3 alc_ ENSG0000025IQCH BMI FAM1C222 ASB9 ARSJ OLFM MFSDEPB4 MCEMr117 ME1 CTNN PC..C36DBF4 DSG2 SOCS2ZNF4 LINC OVCHZNF6 AF28 GGTP PC..C38.1ZMAT r752 SETDr169 r731 CILP OTOF LOC105375r930 ZNF3r102 FBXO36OCLN ENSG0000027SOCS3 r124 E2F2r150 BTG1SKIL TP63C16 PLSCSIGL DBP r657 IRS2 C202 SPTAr482 r229 CLNKBCL2 r360 PPP1R14CCDC17 PC..C44.6KLHD Gly Glc0 AMPDr490 Pro RMRP lPC..C17 ENSG000001H1 ADAMHbA1 Glcs C8.1 NUP2 NPRL PTPR PC..C34.0VANG HHEXr585 PA__Val PPP2 PC..C40.6AC007 ABCADb_S PC..C38.3TwGl ABCGHDL TGLP OGIS NOTCPLD4 AST TNFR IGFBP2MYLI TSKS PC..C32IGFBP1 CXCLTG FLT3 WastInsl ALT Clns BISRTwIn
Ensemble Feature Selection
0.0 0.2 0.4 0.6 0.8 1.0 Median P_cor S_cor LogReg ER_RF Gini_RF AUC_CF ER_CF 0.01 0.010.02 0.020.02 0.020.02 0.02 0.020.02 0.020.02 0.03 0.030.03 0.03 0.03 0.03 0.030.03 0.03 0.03 0.030.03 0.03 0.030.04 0.04 0.04 0.04 0.040.04 0.04 0.04 0.04 0.040.04 0.04 0.04 0.040.04 0.04 0.040.04 0.04 0.050.05 0.05 0.050.05 0.050.05 0.050.05 0.060.06 0.06 0.06 0.060.06 0.06 0.06 0.060.06 0.06 0.06 0.06 0.07 0.07 0.070.07 0.07 0.070.07 0.070.07 0.080.08 0.08 0.080.08 0.080.08 0.080.09 0.09 0.090.09 0.09 0.09 0.09 0.09 0.09 0.090.09 0.090.09 0.09 0.090.10 0.10 0.100.10 0.100.10 0.10 0.110.11 0.11 0.11 0.110.11 0.11 0.110.12 0.12 0.120.12 0.120.12 0.130.13 0.13 0.130.13 0.13 0.130.14 0.14 0.140.14 0.14 0.14 0.14 0.150.15 0.150.15 0.15 0.160.16 0.160.16 0.17 0.170.17 0.170.17 0.18 0.18 0.180.19 0.190.19 0.200.20 0.210.21 0.23 0.23 0.230.24 0.240.25 0.250.26 0.260.26 0.260.27 0.280.28 0.300.31 0.310.33 0.350.39 0.400.41 0.410.44 0.49 0.68 0.79 TwoInsulin BasalISR Insulin ALT Waist Clins FLT3 TG CXCL1 (HPA052502) IGFBP1 (HPA046972) PC.aa.C32.1 TSKS (HPA045729) MYLIP IGFBP2 (HPA004754) TNFRSF8 (HPA042129) AST PLD4 NOTCH2 (HPA064511) OGIS TotGLP1min0 Importance CCR2_HPA065041 PC.aa.C32.1 NOTCH2_HPA064511 TotGLP1min0 TNFRSF8_HPA042129 IGFBP2_HPA004754 AST MYLIP Waist CXCL1_CXCL2_CXCL3_HPA052502 IGFBP1_HPA046972 FLT3 BMI ALT TG Clins TwoInsulin OGIS BasalISR Insulin 50 60 70 80 90 100 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● CXCL1 (HPA052502) IGFBP1 (HPA046972) IGFBP2 (HPA004754) NOTCH2 (HPA064511) CCR2 (HPA065041) TNFRSF8 (HPA042129)
HDL Bilir ubinDir Bilir ubin Chol LDL Liv erIron Alb umin SBP DBP Liv
erInflam HbA1c Glucose
AST GGTP AL T TG Liv erF at VA T W aist BMI SA T TAT TAT SAT BMI Waist VAT LiverFat TG ALT GGTP AST Glucose HbA1c LiverInflam DBP SBP Albumin LiverIron LDL Chol Bilirubin BilirubinDir HDL 0 0.4 0.8 Value 0 150 Color Key and Histogram Count OGISClinsHDL GlucoseSens P ancIronPancF at Liv erIronGGTP ASTAL T HbA1c Glucose Tw oGlucose SBP DBP Glucagonmin0TotGLP1min0 SA T
TAT BMIWaistTG
Liv erF at VA T Tw oInsulin Insulin BasalISR BasalISR Insulin TwoInsulin VAT LiverFat TG Waist BMI TAT SAT TotGLP1min0 Glucagonmin0 DBP SBP TwoGlucose Glucose HbA1c ALT AST GGTP LiverIron PancFat PancIron GlucoseSens HDL Clins OGIS 0 0.4 0.8 Value 0 150 Color Key and Histogram Count
≥ .
β
β
/
−
−
−
−
×
−
≥ .
Important Avgerage bootstrap Simple with MB Both VAT SAT TAT LiverFat Waist HbA1c Glucose Insulin TwoInsulin BasalISR OGIS Clins TotGLP1min0 HDL TG ALT AST GlucoseSens LiverIron BMI GGTP TwoGlucose Glucagonmin0
Strength & direction >= 0.8
SAT VAT TAT LiverFat DBP Waist ALT Albumin GGTP HbA1c HDL BMI TG SBP AST