Generating a Mechanism-Based Taxonomy for AD and PD Using Big Data
1. Mission
To
increase
knowledge
of
the
causes
of
Alzheimer´s
and
Parkinson´s
Disease
by
genera<ng
a
mechanism-‐based
taxonomy;
to
validate
the
taxonomy
in
a
prospec<ve
clinical
study
that
demonstrates
its
suitability
for
iden<fying
pa<ent
subgroups
(based
on
discrete
disease
mechanisms);
to
support
future
drug
development
and
lay
the
founda<on
for
improved
iden<fica<on
and
treatment
of
pa<ent
subgroups
currently
classified
as
having
AD
or
PD.
AETIONOMY:
A
Big
Data
Approach
for
the
Genera:on
of
a
Mechanism-‐Based
Taxonomy
for
AD
and
PD
AD/PD
conference,
March
,
2014
Duncan McHale
Martin Hofmann-Apitius
2.
In
2011,
Kola
and
Bell
published
a
remarkable
paper
in
Nature
Reviews
Drug
Discovery.
With
their
“Call
to
reform
the
taxonomy
of
human
disease”
they
proposed
a
new,
mechanism-‐
based
classifica:on
of
human
disease.
Kola,
I.,
&
Bell,
J.
(2011).
A
call
to
reform
the
taxonomy
of
human
disease.
Nature
Reviews
Drug
Discovery,
10(9),
641-‐642.
3. Add
your
logo
Cartoon
kindly
provided
by
Reinhard
Schneider,
LCSB
5. The
AETIONOMY
concept
is
based
on
a
BIG
DATA
paradigm:
• Comprehensive
and
systema:c
harves:ng,
re-‐annota:on
and
cura:on
of
all
relevant
public
data
• GeneraZon
of
semanZc
frameworks
that
support
the
integra:on
and
interoperability
of
data
through
metadata
annota:on
• SystemaZc
capturing
of
relevant
knowledge
and
modelling
of
disease
in
dedicated,
knowledge-‐based
disease
models
• IdenZficaZon
of
testable
hypotheses
represenZng
putaZve
disease
mechanisms
through
data-‐
and
knowledge-‐driven
model
valida:on
and
mining
• ValidaZon
of
a
selecZon
of
testable
hypotheses
in
the
course
of
a
prospec:ve
clinical
study
The
AETIONOMY
Concept
6. • Organising
all
relevant
data
with
proper
curaZon
and
annotaZon
in
an
indicaZon-‐specific
data
cube
• ConstrucZon
of
disease
models
that
capture
and
represent
available
knowledge
and
make
it
amenable
for
mining
purposes
• Development
of
disease
ontologies
providing
the
semanZc
framework
for
proper
metadata
–
annotaZon
of
data
and
supporZng
informaZon
retrieval
and
knowledge
discovery
• ImplementaZon
of
these
modules
in
a
service-‐oriented
architecture
linking
data
and
knowledge
to
appropriate
data
analysis,
data
visualisa:on
and
knowledge
discovery
services
The
Problem-‐Solving
Approach
in
AETIONOMY
7. The
AETIONOMY
Knowledge
Base:
Organising
Data,
Models
and
Knowledge
to
make
them
amenable
for
modelling
and
mining
Causal Relationship ModelsData Cube (the „axis model“)
Taxonomies
(ordering principles; metadata)
Analysis workflows /
visualisation
14. The
Alzheimer
Disease
Ontology
(ADO)
Malhotra,
Ashutosh,
et
al.
Alzheimer's
&
Demen1a
(2013)
15. Preparatory
Work:
The
Parkinson´s
Disease
Map
PD
Disease
Map
generated
in
CellDesigner
Contributed
by
AETIONOMY
partner
LCSB
(Luxembourg)
See:
h`p://
hdl.handle.net/
10993/2261
16. Capturing
Knowledge
on
Causes
and
Effects:
OpenBEL
a(CHEBI:corticosteroid) -| bp(NCI:”Tissue Damage”)
The abundance of molecules
designated by the name
“corticosteroid” in the CHEBI
namespace.
The biological process
designated by the name
“Tissue Damage” in the NCI
namespace.
decreases
Term
Expression
RelaZonship
Term
Expression
Subject
Predicate
Object
17. The
World´s
largest
Computable
Knowledge
Representa:on
on
AD
Kodamullil,
et
al.
Alzheimer's
&
Demen1a,
in
press
19. OpenBEL
–
based
Mechanism-‐Iden:fica:on
!
OpenBEL
model-‐model
comparison
results
in
the
first
mechanism-‐
hypothesis
generated
in
AETIONOMY:
a
possible
involvement
of
the
NGF-‐
NGFR-‐BDNF
pathway
in
early
decision-‐making
of
the
neuron
on
Neuron
Survival
vs.
Apoptosis.
Note
the
integra:on
of
gene:c
variance
informa:on
in
OpenBEL
Kodamullil
et
al.,
Alzheimer´s
&
DemenZa,
in
press
20. OpenBEL
–
based
Mechanism-‐Iden:fica:on
Mining
of
co-‐morbidity
informaZon
results
in
the
second
mechanism-‐
hypothesis
generated
in
AETIONOMY:
a
possible
link
between
insulin
receptor
pathway,
mTOR-‐induced
autophagy
and
APP
pepZde
clearance
Suppor:ve
evidence
from
SNPs
that
are
shared
by
AD
and
T2DM
!
21. SNP-‐based
iden:fica:on
of
candidate
mechanisms
Contributed
by
Luc
Canard,
Gordon
Ball,
Jesper
Tegner,
Ma`
Page,
Suhas.Vasaikar
23. A
knowledge
base
comprising
curated,
re-‐annotated
and
well-‐organised
data;
knowledge
and
disease
models
that
will
sustain
over
the
next
10
years
Modelling
and
Mining
Workflows
suited
for
the
generaZon
of
new
mechanism-‐hypotheses
with
a
special
focus
on
early
dysregulaZon
events
A
set
of
testable
hypotheses
on
disease
mechanisms
of
which
a
subset
has
been
validated
in
a
clinical
study.
The
clinical
study
links
us
to
the
EPAD
project
A
first
version
of
a
mechanism-‐based
taxonomy
for
AD
and
PD
providing
the
high
level
structure
for
a
future
taxonomy
of
neurodegeneraZve
diseases
that
will
be
populated
with
more
and
more
disease
mechanisms
over
Zme
AETIONOMY
Expected
Outcome
24. N O
M Y
The
People
and
the
InsZtuZons
behind
the
Project
25. AETIONOMY
Partners
Add
your
logo
UCB
EMC
BI,
Fraunhofer,
LUH,
UKB
ICM,
SARD
IDIBAPS
UCL
NeuroRad
PHI
AE,
LCSB
(UL)
KI
Novar:s