webinar : predicting pharmacology and safety profiles with aurpass
DESCRIPTION
Webinar by Aureus Sciences : "Predicting Pharmacology and Safety Profiles with AurPASS"TRANSCRIPT
Make the right decisions with data you can trust
2AurPASS
Instructions for attendees
Due to the High number of attendees:
Before starting Use your full name for Identification
We will mute your microphone during the webinar
During the webinarUse the chat for questions
At the End You can ask your questions orally
We will answer them either in the webinar session or later on by email.
07/04/2011
Predicting Pharmacology & Safety Profiles Using AurPASS
4AurPASS
Aureus Sciences
Knowledge-Based Pharmacological Profiling
AurPROFILER: visualizing the Known
Filling the Gap using AurPASS to Predict the Unknown
AurPASS
PASS Predictive Algorithm
Added Value of AurSCOPE Knowledgebases
AurPASS Ion Channels & AurPASS Kinase
Live Demo: Applications for Safety Pharmacology Assessment
Discussion
Agenda
07/04/2011
5AurPASS
Aureus at a Glance
Aureus is a knowledge and information management solutions provider for the Life Science industry located in Paris.
Over the past 10 years the company has developed a unique knowledge production platform which stores, indexes and organizes critical chemical and bioactivity information including experimental in vitro and in vivo protocols from the public or private literature.
Aureus employees 20 high level scientists (PhD and MSc) including scientific and IT project leaders and documentation analysts
Aureus is funded by institutional investors like CDC Entreprise, AXA Provate Equity, OTC Asset Management and is part of the OSEO Excellence innovation team.
07/04/2011
Aureus Approach
7AurPASS
AurSCOPE Knowledge Databases Target Pharmacological Space
2002 2004 2005 2008 2009
Aureus Terminology, Glossary, Thesaurii>70 glossaries and dictionaries, 5-10 taxonomies, >65,000 terms and synonyms, IUPHAR,
SwissPROT
GPCR
BiologicalProtocols
Unique Ligands
BiologicalActivities
PublicationsAnalyzed
KinaseIon ChannelNuclear Receptor
Protease
146,426354,323
670,407 1,243,793
10,206 including 6,806 patents
13,303 including 6,966 patents
32,588 including 15,281 patents
4,668 including2,484 patents
98,439
275,044
4,226 including1,060 patents
36%
23%16%
13%8% 2% 1%
Activities by Protocol Types
BindingIn vivoElectrophys.FluxIsol. OrganCell Behav.Others
1%2%7%13%
77%
Activities by Protocol Types
FluxBindingIn VivoCell behaviourEnzymology
54%
17%
15%7%
3% 3%1%
Activities by Protocol Types
BindingSecond messengerIn vivoIsolated organCell behaviourFluxEnzymology
418,202
124,313
66%20%
7%5% 2%
Activities by Protocol Types
Enzymology
Cell proliferation
In Vivo
Binding
Flux
186,047
32%
37%
11%
9%6%
2%2%
Activities by Protocoles types
BindingTransactivationIn VivoCell proliferationInductionFluxProtein interaction
48,500
07/04/2011
8AurPASS
AurSCOPE Knowledge Databases ADME/DDI & TOX
BiologicalProtocols
Unique Ligands
BiologicalActivities
PublicationsAnalyzed
hERGADME/DDI Hepatotox
Aureus Terminology, Glossary, Thesaurii>100 glossaries and dictionaries, 5-10 taxonomies, >65,000 terms and synonyms,
IUPHAR, SwissPROT
13,88734,872 compounds including 4,050 metabolites
33,210462,785
2,291 including 547 Patents
12,165 including200 FDA Reviews
3,600 Articles (67 patents, 75 FDA Studies)
70,035
845 Drugs
1% 1%3%
4%28%
61%
Activities by Protocol Types
BiochemistryCell behaviourBindingInductionFluxEnzymologyIn VivoOthers
67%
14%
13%
4% 1% 1%
Actvities by Protocol Types
PK / DDIMetabolism (Enzyme)Inhibtion (Enzyme)TransportInductionBinding
53%
27%
10%10%
Activities by Protocol Types
ElectrophysiologyIn VivoBindingFlux
07/04/2011
Make The Right Decision With Data You Can Trust
AurQUESTData mining
AurPROFILERData analysis & NavigationMedicinal Chemistry Space
AurPASSPredict Biological Activities SpectraDDI Predict
AurPASS
10AurPASS
AurPROFILER – Pharmacology Profiles of 74 Drugs (Big Pharma)
07/04/2011
11AurPASS
Filling the Gap Using AurPASS in silico Predictions
07/04/2011
Aureus
Prediction of
Activity
Spectra for
Substances
12AurPASS
Predicting New Activities on Selected big Pharma Drugs
277 Ion Channels
86
Sel
ect
ed
Mol
ecu
les
fro
m M
ER
CK
07/04/2011
13AurPASS
Over Forty Publications with Independent Confirmation of PASS Predictions…
07/04/2011
14AurPASS
Training SetSAR data extracted from individual AurSCOPE databases
Chemical & biological post-processing
Chemical StructuresMultilevel Neighborhoods of Atoms (MNA) descriptors
SAR BaseDiverse chemical structures and associated MNA descriptors
Associated qualitative biological activity types
Prediction AlgorithmPASS Bayesian approach
Models ValidationLOO and L20%O validations
External dataset
AurPASS
07/04/2011
15AurPASS
Training SetSAR data extracted from individual AurSCOPE knowledgebases
Chemical & biological post-processing
Chemical StructuresMultilevel Neighborhoods of Atoms (MNA) descriptors
SAR BaseDiverse chemical structures and associated MNA descriptors
Associated qualitative biological activity types
Prediction AlgorithmPASS Bayesian approach
Models ValidationLOO and L20%O validations
External dataset
AurPASS
07/04/2011
16AurPASS
Added Value of Structured Data for Predictive Models
07/04/2011
Chemical Processing
Biological Processing
• No Peptides• No mixtures• Inorganic and metalo-organic removal• Charge standardization
• Biological protocols• Target hierarchy • Target species, cell lines and tissues• Biological activity parameters• Ligand action on target
17AurPASS 07/04/2011
Assigning Activity Classes
AurPASS Activity Types
18AurPASS
Multilevel Neighbourhoods of Atoms (MNA) descriptors are generated on the basis of the connection table of 2D molecular structures (including hydrogens)
Bond types are not specified
MNA are generated as recursively defined • Zero level MNA descriptor for each atom is the atom type itself;
• The first level MNA descriptor include the atom’s zero level descriptors and zero-level descriptors of its neighboring atoms sorted lexicographically …
Encoding Chemical Structures: MNA Descriptors
07/04/2011
MNA descriptors dictionary
19AurPASS
Encoding Chemical Structures: MNA Descriptors
07/04/2011
MNA/0: C
MNA/1: C(CN-H)
MNA/2: C(C(CC-H)N(CC)-H(C))
CC
H
C
O
O
NC
H
C
CH
H H
CC
H
C
O
O
NC
H
C
CH
H H
CC
H
C
O
O
NC
H
C
CH
H H
MNA/2
C(C(CC-H)C(CC-C)-H(C))C(C(CC-H)C(CN-H)-H(C))C(C(CC-H)C(CN-H)-C(C-O-
O))C(C(CC-H)N(CC)-H(C))C(C(CC-C)N(CC)-H(C))N(C(CN-H)C(CN-H))-H(C(CC-H))-H(C(CN-H))-H(-O(-H-C))-C(C(CC-C)-O(-H-C)-O(-C))-O(-H(-O)-C(C-O-O))-O(-C(C-O-O))
20AurPASS
PASS Approach for Biological Activity Prediction
07/04/2011
According to the Bayes' theorem, the probability P(A|S) that the compound S has activity (or inactivity) A, equals to:
P(A|S) = P(S|A)•P(A)/P(S)
If the descriptors of organic compound D1, ..., Dm are independent, then:
P(S|A) = P(D1, ..., Dm|A) = ПiP(Di|A)
P(A) and P(A|Di) are calculated as sums through all compounds of the training set:
k ik
k kiki )(Dg
(A))w(Dg)|DP(A
i k ik
i k kik
)(Dg
(A))w(DgP(A)
Filimonov D.A., Poroikov V.V. (2008). Probabilistic Approach in Virtual Screening. In: Chemoinformatics Approaches to Virtual Screening. Alexander Varnek and Alexander Tropsha, Eds. RSC Publishing.
21AurPASS
The list of activities which are probable for a particular compound with the estimates for each activity of :
Pa: probability to be active
Pi: probability to be inactive
Pa and Pi are calculated independently: Pa + Pi 1
Pa (Pi) can be considered as the probability of the compound belonging to classes of active (inactive) compounds
PASS Approach for Biological Activity Prediction
07/04/2011
22AurPASS
AurPASS Ion Channels – Version 1.2.1
07/04/2011
Training Set47 938 molecules517 Activity Types
Mean Accuracy of Prediction: 98%
Test Set2244 molecules113 Activity types
Mean Accuracy of Prediction: 90%
Prediction Accuracy using AurPASS
0
20
40
60
80
1005HT-3
AMPA
AMPA/Kainate
ASIC
Ca2+
CFTR
ENaC
GABAA
InsP3
K+KainateKATP
L-type Ca2+
Na+
nACh
NMDA
N-type Ca2+
P2X
TRPA
TRPV
T-type Ca2+AurPASS Activity
IAP
24AurPASS
How AurPASS Works?
07/04/2011
Ion ChannelsKinasesGPCRsProteasesNuclear ReceptorsOff-Targets
25AurPASS
AurPASS Ion Channels
07/04/2011
Live Demo
26AurPASS
AurPASS Product Line
07/04/2011
May, 2011AurPASS Off-TargetsAurPASS GPCRAurPASS Nuclear Receptors
Application Note and Evaluation Software
available Under Request
27AurPASS
Thank you for your
attention
Aureus Sciences 174, Quai de Jemmapes 75010 Paris, FRANCE
www.aureus-pharma.com
07/04/2011