Exploring available compound data with the Open PHACTS Discovery Platform and KNIME

252nd ACS National Meeting

Daniela Digles, Gerhard F. Ecker Philadelphia, PA, August, 21, 2016

Pre-competitive Informatics:Pharma are all accessing, processing, storing & re-processing external research data

LiteraturePubChem

GenbankPatents

DatabasesDownloads

Data Integration Data AnalysisFirewalled Databases

Repeat @

each

company

x

Lowering industry firewalls: pre-competitive informatics in drug discovery

Nature Reviews Drug Discovery (2009) 8, 701-708 doi:10.1038/nrd2944

Different concept types

@gray_alasdair Big Data Integration 4

Nanopub

Db

VoID

Data Cache (Virtuoso Triple Store)

Semantic Workflow Engine

Linked Data API (RDF/XML, TTL, JSON)Domain

Specific

Services

Identity

Resolution

Service

Chemistry

Registration

Normalisation

& Q/C

Identifier

Management

Service

Indexing

Co

re P

latf

orm

P12374

EC2.43.4

CS4532

“Adenosine

receptor

2a”

VoID

Db

Nanopub

Db

VoID

Db

VoID

Nanopub

VoID

Public Content Commercial

Public

Ontologies

User

Annotations

Apps

Workflow tools

9Daniela Digles

• Single „blocks“ for each data processing step(e.g. data reader, calculations, visualization, …)

• Blocks are placed via drag-and-drop andconnected to each other with arrows.

• Commercial (e.g. Pipeline Pilot) and free tools(e.g. KNIME) available.

Data inputData

processingData

processing

Data export

View data

KNIME

10Daniela Digles

• KNIME Analytics Platform

• Available from www.knime.org

• Open source data analytics, reporting andintegration platform

• Workflows can be built by connecting „Nodes“

• Open PHACTS KNIME nodes available fromgithub: https://github.com/openphacts/OPS-Knime

Open PHACTS KNIME

11Daniela Digles

executable API call

Swagger

12Daniela Digles

Structured format for the generation of API documentation.https://dev.openphacts.org/swagger/spec/ops_1_5.json

(….)

Swagger

13Daniela Digles

https://dev.openphacts.org/docs/

Open PHACTS KNIME

14Daniela Digles

Open PHACTS KNIME

15Daniela Digles

Obtaining the results

16Daniela Digles

Answering “scientific competency questions”

17Daniela Digles

• 20 questions defined at the beginning of theproject.

• Example: Give me all oxidoreductase inhibitors active <100 nM in human and mouse.

• Many questions need a combination of queriesto the Open PHACTS Platform.

Questions: Azzaoui K et al. (2013) Drug Discov. Today 18: 843 – 852.

Workflows: Chichester C et al. (2015) Drug Discov. Today 20: 399 – 405.

Example workflow

18Daniela Digles

Q10: For a given compound, summarize all similar compounds and their activities

CC1=C(C(C(=C(N1)C)C(=O)OC)C2=CC=CC=C2[N+](=O)[O-])C(=O)OC

Workflow to collect compound data

19Daniela Digles

• Data for retrieved molecules:• Function and toxicity annotation (Drugbank)

• Role of the molecule (ChEBI)

• Pharmacology data, activity < 10 µM (ChEMBL)

• Patent data (SureChEMBL)

• Data for retrieved targets:• Pathways (WikiPathways)

• Diseases (DisGeNET)

• Example: propafenone derivative

Workflow to collect compound data

20Daniela Digles

Collected results:Compound information

21Daniela Digles

• Structure search: 96 molecules, including themolecule itself.

• Compound information/classification: 1 knowndrug propafenone

Collected results:Patent information

22Daniela Digles

• Highest confidence score:• Patents found for 3 molecules

• No patents found for the original structure

• Lower confidence score:• Patents found for 8 molecules

• 2 patents found for the original structure (High throughput assay for discovering new inhibitors of the GIRK1/4 channel)

• Restriction: Markush structures are not enumerated in SureChEMBL

Collected results:Bioactivity values

23Daniela Digles

191 activity values (lower than 10 µM) against 33 targets.

P-glycoprotein

Cells expressing P-glycoprotein

Propafenone

Collected results:Targets

24Daniela Digles

Target classifications (per compound)

Target classifications (unique targets)

Collected results:Pathways for targets

25Daniela Digles

• 98 Pathways in total

• 4 pathways contain > 5 of the identified targets• GPCR downstream

signalling

• GPCR ligand binding

• Relevance of the pathways?

Collected results:Diseases for targets

26Daniela Digles

> 2000 diseases in 25 disease classes

• Workflow allows the easy preparation of a first overview on known data for a compound of interest.

• New ideas for targets to test the compounds against. • Example: Serotonin receptor for propafenone

derivatives

• Literature (Pubmed) is returned for the results.

• Additional external or in-house data can be added.

• Methods for prioritization needed:• Relevance of pathways

• Relevance of diseases

Daniela Digles 27

Conclusions

Useful links

28Daniela Digles

Open PHACTS: http://www.openphactsfoundation.org/

API: https://dev.openphacts.org/

Support portal: http://support.openphacts.org/

Example Workflows: http://www.myexperiment.org/groups/1125.html

Presentations on YouTube:

https://www.youtube.com/user/OpenPHACTS

For help or feedback: support@openphacts.org

Acknowledgements

Pharmacoinformatics research group, University of Vienna

– Gerhard F. Ecker

– Barbara Zdrazil

– Jana Gurinova

Open PHACTS – KNIME

– Ronald Siebes, VU Amsterdam

– Christine Chichester, SIB

– Evan Tzanis, QMUL

Daniela Digles