2013 dec bgu_israel_luciano_dec_22

1

Semantic eHealth: Getting more out of biomedical data using

Semantic TechnologyInstructors:

Joanne S. Luciano, PhDRensselaer Polytechnic Institute, University of California, Irvine, USA

Eitan Rubin, PhDBen-Gurion University

December 22-25, 2013

Ben-Gurion University of the Negev, Israel

2

Instructor

Lecturer,

Department of Microbiology and Immunology

Faculty of Health Sciences

InterestsUnderstand the role genetics plays in the development of diseases

Research

Novel methods for disease stratification using genetic analysis as predictors of treatment outcomes.

Improved methods for computational target prioritization in genetic association studies

An end-user programming language for biologists

Email: [email protected]

3

Instructor

Joanne S. LucianoDeputy Director

Web Science Research Center

Interests

Use and Develop Technology. Infrastructure and Analytics to Advance Science and Increase its Utility to Improve Health Outcomes

Research

BioPAX, TMO, InfluenzO

General Framework for Ontology Evaluation

Systems Biology and Medicine - Major Depressive Disorder (MDD)

Medicine, Health, WellbeingEmail: [email protected]

4

Timeline(earlier work: 10 years in Software Research & Development and Product Development)

20091993

World Congress on Neural Networks, July 11-15, 1993, Portland, Oregon SIGMental Function andDysfunctionSam Levin

Jackie Samson,Mc Lean HospitalDepression Research

1996

1995

20081994

Patents Soldto Advanced

Biological Laboratories

Belgium

Patents Offered at

Ocean Tomo Auction

Chicago, IL

US Patent No. 6,317,73

Awarded

US PatentsNo.

6,063,028Awarded

2001

2000

PhD

Thesis Proposal Approved

Workshop Neural Modeling of Cognitive and Brain Disorders

BioPAX

?Linked DataW3C HCLSBioDASH

EPOS

2006

EMPWR

Poster Presented ISMB 1997PSB 1998

1997

2010

Rensselaer(RPI)

20112012

2013

U PittGreg Siegle

Collaboration

YuezhangXiao

Master’sThesis(RPI)

Brendan AshbyMaster’sThesis

(RPI)

Center forMulti-

disciplinary Research

andDepressionTreatmentSelection

5Predictive Medicine, Inc. © 20105

Overview

Promises:

0. Introduction – Depression ResearchHow did a nice girl like me,

wind up in a field like this?

1.Intro to Data Science

2.Tools to Integrate Biomedical Data

3.Knowledge Standards and Best Practices that enable web scale Integration


BioPathways Consortium

BioPAX

W3C Semantic Web for Health Care and Life Sciences (HCLSIG)

Establishing Communities of Interest/Practice

7

BioPAX - Enabling Cellular Network Process Modeling

MetabolicPathways

MolecularInteractionNetworks

SignalingPathways

Gene RegulatoryNetworks

Glycolysis Protein-Protein Apoptosis TFs in E. coli


Translational Medicine

• Rapid transformation of laboratory findings into clinically focused applications

• ‘From bench to bedside and back’• “and back” includes patients!


HUGE PROBLEM

Characterized by persistent and pathological sadness, dejection, and melancholy

Prevalence (US)

6% year (18 million)

16% experience it in their lifetime

Cost

44 Billion (1990)

Impact

1% Improvement means (180, 000 people helped)

1% Improvement means (440 million in savings)


Widespread


Treatment Choice VagueNo easy answer


Overview

• Why we did this work - to improve quality of life for millions of people suffering from depression

• How we did it - used differential equations (“neural network”) to model and compare response to different antidepressant treatments

• What we found - different response patterns for the two treatments - the order and timing of improvement of symptoms were different

• What we think it means - improvement in selection of treatment thereby reducing unnecessary costs and suffering. Potentially saving lives

1313

Research Goals

Illuminate recovery course

(personalized)

Properly diagnose and properlymatch patient with the best individualized

treatment option available, includingnon-drug treatments


Today’s talk focuses on:Response to treatment

Treatment Response Study


Depression Background

• Clinical Depression

• Treatment

• Symptom Measurement

• No specific diagnosis

• No specific treatment


Clinical Data

Symptoms

-HDRS (0-4 scale)

Treatment-Desipramine (DMI)

-Cognitive Behavioral Therapy (CBT)

Outcome

- Responders


Hamilton Psychiatric Scale for Depression

Clinical Instrument standard measure in clinical trials. Example of first three items of 21 items that measure individualSymptom intensity.


Why Model?

Easier to understandEasier to manipulateEasier to analyze

Recasting the problem into mathematical termsmakes it:


Understanding Recovery


Understanding Recovery


Depression Data

7 Symptom Factors

Physical: E Sleep M, L Sleep

EnergyPerformance: Work & InterestsPsychological: Mood

CognitionsAnxiety

2 Treatments Cognitive Behavioural Therapy (CBT)Desipramine (DMI)

Clinical Data Responders = improvement >= 50% on HDRS totalN = 6 patient each study

6 weeks = 252 data points (converted to daily)

each study (CBT and DMI)


Overview Recovery Model and Parameters

M

EW

MS

ES

A

C


Recovery Equation(Luciano Model)

+

+

+

-==


Individual Patient Recovery Pattern and Error

Example Patient (CBT)

Fit of Model on for individual patient captures trends but not entire pattern. Not good enough.


Patient Group (CBT)

Recovery Pattern and Error

Model on data for patient treatment group captures entire pattern. Good fit of Model to data.


Latency


Treatment Effects and Interaction Effects

CBTSequential

DMI(delayed)

CONCURRENT

DMI: •Interactions > 2x •Loops


Order and Time a symptom improves are both different

Different Response Patterns for Different Treatment

CBT DMICBT (talk: no drugs) DMI (drug: tricyclic antidepressant)

This is important because it shows how an antidepressant medication could lead to a suicide.

By giving a suicidal patient DMI, you could increase the patients energy before the suicidal thoughts improve. This could give them the energy to act on those suicidal thoughts.


Overview

• Why we did this work - to improve quality of life for millions of people suffering from depression

• How we did it - used differential equations (“neural network”) to model and compare response to different antidepressant treatments

• What we found - different response patterns for the two treatments - the order and timing of improvement of symptoms were different

• What we think it means - improvement in selection of treatment thereby reducing unnecessary costs and suffering. Potentially saving lives.


Give me a break!!!Give me a break!!!

One more slide One more slide (so you see what’s coming (so you see what’s coming

whenwhenwe return)we return)


Inside the Overview

1. Intro to Data ScienceShifts (programs to data, populations to individuals, hoarding to sharing)

What makes data useful?

Can we exploit the web to access data?

2. Tools to Integrate Biomedical DataBy Hand

Using Tools

Automated

3. Knowledge Standards and Best Practices that enable web scale Integration

Connecting data

5 Stars

5 Stars not enough


Give me a break!!!Give me a break!!!


Inside the Overview

1. Intro to Data ScienceShifts (programs to data, populations to individuals, hoarding to sharing)

What makes data useful?

Can we exploit the web to access data?

2. Tools to Integrate Biomedical DataBy Hand

Using Tools

Automated

3. Knowledge Standards and Best Practices that enable web scale Integration

Connecting data

5 Stars

5 Stars not enough

34

Intro to Data Science

What do you think data is?

What could data science possibly mean?

Can data be reused once the original purpose (study) is done?

Predictive Medicine, Inc. © 2010

35

Data, Not Programs

351. Webopedia. “Data Dictionary.” Available online at www.webopedia.com/TERM/d/data_dictionary.html.

12

3636


Feet?Years?December?Noon?Dozen?

Data, Not Programs


12Feet?Years?December?Noon?Dozen?

37

Data, Not Programs


38

Data, Not Programs

Data Dictionaries:

Without a data dictionary, a database management system [or any program] cannot access data from the database.”1


Duh!

39

Data, Not Programs

Data Dictionaries:

Without a data dictionary, a database management system [or any program] cannot access data from the database.”1


Duh!

40

Metadata (simplified)

40

Biochemical Reaction

<reaction id=“pyruvate_dehydrogenase_rxn”/>

<listOfReactants> <speciesRef species=“NADP+”/> <speciesRef species=“CoA”/>

<speciesRef species=“pyruvate”/>

</listOfReactants> <listOfProducts> <speciesRef species=“NADPH”/> <speciesRef species=“acetyl-CoA”/> <speciesRef species=“CO2”/> </listOfProducts> <listOfModifers> <modifierSpeciesRef

species=“pyruvate_dehydrogenase_E1”/>

</listOfModifiers>

</reaction>

Synonyms

<species id=“pyruvate” metaid=“pyruvate”><annotation xmlns:bp=“http://biopax.org/release1/biopax_release1.owl”/><bp:smallMolecule rdf:ID=“#pyruvate” > <bp:SYNONYMS>pyroracemic acid</bp:SYNONYMS> <bp:SYNONYMS>2-oxo-propionic acid</bp:SYNONYMS> <bp:SYNONYMS>alpha-ketopropionic acid</bp:SYNONYMS> <bp:SYNONYMS>2-oxopropanoate</bp:SYNONYMS> <bp:SYNONYMS>2-oxopropanoic acid</bp:SYNONYMS> <bp:SYNONYMS>BTS</bp:SYNONYMS> <bp:SYNONYMS>pyruvic acid</bp:SYNONYMS></bp:smallMolecule></annotation></species>

41

Instead of textual labels <bp:smallMolecule rdf:ID=“#pyruvate”> <bp:Xref> <bp:unificationXref rdf:ID=“#unificationXref119"> <bp:DB>LIGAND</bp:DB> <bp:ID>c00022</bp:ID> </bp:unificationXref> </bp:Xref> </bp:smallMolecule>

Use actual URIs

Metadata (Webified)

41

42

Query results return

links to the original data!

Metadata (Webified)

Adapted from Mark Wilkinson webscience20-120829124752-phpapp01

43

Data Sharing (Shafu)



Had enough for now?Had enough for now?

Ready to start getting your Ready to start getting your hands dirty?hands dirty?

45

Joanne S. Luciano, BS, MS, PhDAcademic:

[email protected]

Rensselaer Polytechnic Institute, Troy, NY

University of California – Irvine, CA

Consulting:

[email protected]

Predictive Medicine, Inc., Belmont, MA


CV Background slides...


Whew!Whew!

Now that was fun, wasn’t Now that was fun, wasn’t it?it?

Any questions?Any questions?


Neural Modeling of Depression

1996 Luciano, J., Cohen, M. Samson, J. ”Neural Network Modeling of Unipolar Depression,” Neural Modeling of Cognitive and Brain Disorders, World Scientific Publishing Company, eds. J. Reggia and E. Ruppin and R. Berndt. Book cover; chapter pp 469-483.

Luciano Model highlighted on book cover

Workshop 1995Book 1996


Inside the Overview

1. Tools to Integrate Biomedical Data• By Hand

• Really by hand, i.e. depression research

• Cutting and pasting between text editors, spreadsheets, and command lines

• Using Tools • KNIME

• Automated • Protégé

• Gruff & Allegrograph

49

Diabetes Classification

WHO Recommendation 2011

HbA1c 48 mmol/mol (6.5%) cut point• stringent quality assurance tests

• assays are standardised to international reference values,

• no conditions present which preclude its accurate measurement.

A value of less than 48 mmol/mol (6.5%) does not exclude diabetes diagnosed using glucose tests.


50

Diabetes Classification

Situations where HbA1c is not appropriate for diagnosis of diabetes:

• ALL children and young people

• Patients of any age suspected of having Type 1 diabetes

• Patients with symptoms of diabetes for less than 2 months

• Patients at high diabetes risk who are acutely ill (e.g. those requiring hospital admission)

• Patients taking medication that may cause rapid glucose rise e.g. steroids, antipsychotics

• Patients with acute pancreatic damage, including pancreatic surgery

• In pregnancy

• Presence of genetic, haematologic and illness-related factors that influence HbA1c and its measurement - see Annex 1 from WHO report


2013 dec bgu_israel_luciano_dec_22

Health & Medicine

recoverypredictive medicine

analyzepredictive medicine

widespreadpredictive

livespredictive medicine

savings predictive medicine

depression data

easy answerpredictive

data science