Transcript Advanced Data-Visualization for Drug Discovery

Using Spotfire DecisionSite to Realize
the Full Value of High-Throughput
Screening ADME Data
Eric Milgram
Pfizer Global Research & Development – La Jolla
Spotfire Users’ Group Meeting
Wednesday October 15, 2003
San Francisco, California
Challenge faced by Pharmaceutical
Industry
$25
 Reduce Attrition
$20
Cost and Number of NDAs per year
$15
 Increase Productivity
$10
 No growth
$5
$0
19
7
19 0
7
19 1
7
19 2
7
19 3
7
19 4
7
19 5
7
19 6
7
19 7
7
19 8
7
19 9
8
19 0
8
19 1
8
19 2
8
19 3
8
19 4
8
19 5
8
19 6
8
19 7
8
19 8
8
19 9
9
19 0
9
19 1
9
19 2
9
19 3
9
19 4
9
19 5
9
19 6
9
19 7
9
19 8
9
20 9
00
 Budgetary Pressure
Source: PhRMA annual survey, 2000
Why Do Candidates Fail?
Commercial Reasons
Adverse effects
in man
Miscellaneous
5%
198 NCEs
5%
10%
39%
Animal Toxicity
11%
30%
Lack of Efficacy
Drug Discovery Today 2:436, 1997
Pharmacokinetics
The Fate of a Medication after Administration
ADME
• Absorption
• The movement of drugs into the bloodstream or
lymphatic system from the site of administration
• Distribution
• The distribution of absorbed drugs from the absorption
site to all areas of the body
• Metabolism
• The biotransformation of drugs to more polar forms
(hydrolysis, oxidation, conjugation, etc.)
• Excretion
• The elimination of “unwanted” substances
Drug Metabolism in Drug Discovery
•
Early assessment is critical, since the duration of action is
dependent on structural modifications induced by in vivo
metabolizing systems.
•
Early knowledge of metabolic products permits
metabolism guided structure modification schemes, such
as modification of metabolic “soft spots” to achieve
prolonged drug action.
• Identify pharmacologically or toxicologically active
metabolites.
Physicochemical & Biochemical
In- Vitro Assays
Predictive of In-Vivo Absorption, Distribution, Metabolism, Excretion
• Solution Properties
•
•
•
•
Solubility
Log D
Protein Binding
pKa
• Absorption
• PAMPA
• Caco-2, MDCK
• Pgp transport
• IAM
• Metabolism
• Metabolic stability
• Liver microsomes, S-9,
hepatocytes
• Metabolic profile
• CYP450 enzyme inhibition
• Safety Assessment
•
•
•
•
Cell viability
Mutagenesis (Ames)
Glutathione level
Dofetilide binding
Pritchard, et al., “Making Better Drugs: Decision Gates in Non-Clinical Drug Development”
Nature Reviews: Drug Discovery, 2003, vol 2(7), pp. 542-553. (http://www.nature.com/reviews/)
Advances in Laboratory Robotics
and Instrumentation Have Been Swift
Difficulties Resulting from HTS

The rate at which we can collect data far
exceeds our capacity to transform this data
into information that can be used most
effectively to drive important business
decisions



Relevance of data?

Unmasking subtleties (ie “data-mining”)
Number of data dimensions?
What do we do when two different dimensions
are in conflict?
Visualization is a Powerful Tool
For Data Analysis
Spotfire can be used to find trends related to how
samples are formatted on plates.
Scatter Plot
PLATE_NUMBER - 1
PLATE_NUMBER - 2
PLATE_NUMBER - 3
PLATE_NUMBER - 4
PLATE_NUMBER - 5
PLATE_NUMBER - 6
PLATE_NUMBER - 7
PLATE_NUMBER - 8
PLATE_NUMBER - 9
PLATE_NUMBER - 10
PLATE_NUMBER - 11
PLATE_NUMBER - 12
PLATE_NUMBER - 13
PLATE_NUMBER - 14
PLATE_NUMBER - 15
PLATE_NUMBER - 16
PLATE_NUMBER - 23
PLATE_NUMBER - 24
H
G
F
E
D
C
B
A
H
G
F
E
D
C
B
A
H
G
F
E
D
C
B
A
H
G
F
E
D
C
B
A
2
4
6
8
10
12
2
4
6
8
10
12
2
4
6
8
WELL_COLUMN
10
12
2
4
6
8
10
12
2
4
6
8
10
12
Internal Standard Signal vs Acquisition Date
IS Peak Area
Liquid Handling
Error
Inconsistent Sample
Handling
Dirty MS Source
17-Apr-2003 3:59:09 PM
20-Apr-2003
Acquisition Date
How do we use Spotfire DecisionSite
to Allocate Resources Efficiently?

Quality Control and Quality Assurance

Results Analysis and Trending
When combined with chemometrics techniques, such as principal
components analysis (PCA), Spotfire enables viewing of
interesting trends in large, multidimensional data sets.
PCA (2 components) For LJ-EDT Data (human/rat liver/microsomes and caco-2 AB/BA)
UNSTABLE RHEP
STABLE HLM
60
40
UNSTABLE RLM
20
STABLE RLM
0
-20
-40
-60
-80
UNSTABLE HLM
STABLE RHEP
-100
-125
-100
-75
-50
-25
PCA 1
0
25
50
75
Spotfire enables viewing of trends that would be
difficult to spot otherwise
IS Vals vs Ret T ime
ASSAY_TYPE
ASSAY_TYPE
140000
120000
100000
80000
60000
40000
20000
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
RET_TIME_MIN
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Summary
 Spotfire DecisionSite enables powerful interrogation of




large data sets
Ability to generate quickly new views of the same
dataset is essential in a high-throughput discovery
environment
Sometimes, weaknesses in experiment design are
uncovered
Having a collection of “standard” visualizations greatly
facilitates QA
Integration of chemometrics tools (e.g. clustering,
PCA, etc) enables researchers to “gain a deeper
understanding of their data” (DataInformation)