View
13
Download
1
Category
Preview:
Citation preview
Building innovative
drug discovery alliances
Knime – Desktop tools for chemists
Evotec AG, Knime – desktop tools for chemists, May 2011
PAGE
Agenda
1
Knime
Getting project data from Excel spreadsheets
Getting project data from IJC databases
Getting public data from databases
Enumerating libraries of compounds
Getting data from other databases
Knime Server
Knime node development
Conclusions
PAGE 2
A KNIME Solution
The Multi-Purpose Tool in the MedChem Toolkit
PAGE
KNIME
The Multi-Purpose Tool in the MedChem Toolkit
3
PAGE
Knime Workbench
4
PAGE
Agenda
5
Knime
Getting project data from Excel spreadsheets
Getting project data from IJC databases
Getting public data from databases
Enumerating libraries of compounds
Getting data from other databases
Knime Server
Knime node development
Conclusions
PAGE
Extracting data from Excel
Loading into IJC
Project compound data is kept on an Excel spreadsheet
Data is filtered and prepared for loading into an IJC database
Additional calculations are carried out during the workflow
6
PAGE
Agenda
7
Knime
Getting project data from Excel spreadsheets
Getting project data from IJC databases
Getting public data from databases
Enumerating libraries of compounds
Getting data from other databases
Knime Server
Knime node development
Conclusions
PAGE
Using InstantJChem databases
Projects keep their data in Oracle
Every project uses InstantJChem for their project data Project members can easily view and query their data
Easy for project members to load new data and maintain old data
Two schemas; one for project data, one for project ideas
Additional properties can be calculated during data loading
Shared Oracle database All data backed up
Each project area is secure
Only project members can see the data
8
PAGE
Overview
9
Shared IJC project database setup at Evotec
Enterprise licenses for server use shared across Abingdon and Hamburg sites
Dedicated IJC server at Abingdon with Oracle 10g and JChem cartridge
One Oracle schema per drug discovery project containing project data (currently 40 live IJC
shared projects)
A project IJC schema is connected to each Oracle schema
Built-in IJC security & roles used for user management
Deployment to project users via web server URLs
PAGE
Project Data Flow
Drug Discovery IJC Project at Evotec
Compound
Registration
Medicinal
Chemistry
ADMET
10g
IJC 5.3.1
JChem 5.3.1
Abingdon
Opera™
Assays
Counter
Assays
In Vivo
Hamburg
10
PAGE
Extract data from IJC
Project ideas database
Project chemists add synthetic ideas to their IJC database
Knime is used to extract the data that has been added that month
The extracted data is presented in PowerPoint
11
PAGE
Data Model
Schema Editor view: Data Trees
SAR data tree contains joined chemistry &
biology tables, allowing many assay results for
each assay type per compound structure
Assay
tables
Structure table fields
Relationships
All Structures
Bioassay ADMET assays
1-n1-n
12
PAGE
Creating SAR tables from databases
Producing reports
Project data can be queried and filtered
easily
Data from multiple tables can be used
Additional properties can be calculated
SAR table reports can be prepared
from the filtered data
13
PAGE
Extracting data from IJC
14
Using activity data
PAGE
Exporting data from multiple tables
15
Easy data extraction from IJC
PAGE
Agenda
16
Knime
Getting project data from Excel
Getting project data from IJC databases
Getting public data from databases
Enumerating libraries of compounds
Getting data from other databases
Knime Server
Knime node development
Conclusions
PAGE
Accessing public data
Viewing ChEMBL data
ChEMBL is a database of
bioactive drug-like small
molecules, it contains 2-D
structures, calculated properties
(e.g. logP, Molecular Weight,
Lipinski Parameters, etc.) and
abstracted bioactivities (e.g.
binding constants, pharmacology
and ADMET data
17
PAGE
Accessing public data
Viewing ChEMBL data through IJC
18
PAGE
Searching the ChemBL database
Easy access to ChEMBl data using Knime
A simple workflow allows users to search the ChEMBl data
Users don’t need a username and password
19
PAGE
Agenda
20
Knime
Getting data from Excel spreadsheets
Getting project data from IJC databases
Getting public data from databases
Enumerating libraries of compounds
Knime Server
Knime node development
Conclusions
PAGE
Library enumeration
An easy way for chemists to create compounds
Chemists can quickly create a virtual library
Calculate physical properties
Select interesting compounds for further processing
21
PAGE
Virtual Library Enumeration: CDK2 Example
22
PAGE
Re-usable components
Physical properties
Parts of workflows can be re-used
Often used operations can be bundled into meta nodes
23
PAGE
Even more meta nodes!
24
PAGE
Library enumeration and searching
Check EVOsource for starting materials
Enumerate a focused library
Select the promising
compounds
Search in-house compound
database for starting
materials
25
PAGE
Agenda
26
Knime
Getting data from Excel spreadsheets
Getting project data from IJC databases
Getting public data from databases
Enumerating libraries of compounds
Getting data from other databases
Knime Server / Sharing Workflows
Knime node development
Conclusions
PAGE
Search Compound database
Retrieving hits from EVOsource
Using a list of part ID’s
Get lists of suppliers that the compound can be bought from
27
PAGE
Administration of the ELN
Auditing users
Connect to ELN database
Carry out audits By project
By user
Regular
Leavers
Submitted, countersigned, Printed
28
PAGE
Agenda
29
Knime
Getting project data from Excel
Getting project data from IJC databases
Getting data from public data sets
Enumerating libraries of compounds
Getting data from other databases
Knime Server / Sharing Workflows
Knime node development
Conclusions
PAGE
Knime Server
30
Sharing workflows
Workflows created by expert users
Reports for other users
PAGE
SharePoint
31
Sharing workflows
Using SharePoint to share workflows and tips
PAGE
Agenda
32
Knime
Getting project data from Excel
Getting project data from IJC databases
Getting data from public data sets
Enumerating libraries of compounds
Getting data from other databases
Knime server / Sharing Workflows
Knime node development
Conclusions
PAGE
EVOnodes
KNIME Development Projects
Have developed KNIME nodes to run linux based CADD programs
33
PAGE
EVOnodes
34
Knime development projects
Knime nodes to replace
Evotec desktop modelling
tool
Allow integrating modelling
tools in workflows
PAGE
Agenda
35
Knime
Getting project data from Excel
Getting project data from IJC databases
Getting data from public data sets
Enumerating libraries of compounds
Getting data from other databases
Knime Server / Sharing Workflows
Knime node development
Conclusions
PAGE
Conclusions
36
Why use Knime?
On every chemists desktop
Easy for chemists to tailor make database exports
Easy for chemists to enumerate libraries
Easy for chemists to generate tailor made reports from databases
Simple for developers and computational chemists to extend
Fun to work with!
PAGE
Acknowledgements
37
eScience Ian Berry
Catherine Reisser
Mike Potterton
Computational Chemistry Mike Mazanetz
Michelle Szeto
Discovery Chemistry Ed Walker
Rich Jones
Chris Stimson
Your contact:
Building innovative
drug discovery alliances
Dr Bob MarmonSenior Applications Developer, eScience
+44.(0).1235.861561 +44.(0).1235.441402bob.marmon@evotec.com
Recommended