Magnus Fontes Lund University and Institut Pasteur

Magnus Fontes

Lund University and Institut Pasteur

PRECISION MEDICINE FOR

PERSONALIZED HEALTHCARE

Personal data

Hp-35 (1972) ABC80 (1978)

16 kB RAM

Sparc Station

Beginning of the 90’s

Up to 500 MB in RAM

Slide rule

A Brief personal history of computing

From Wikimedia Commons

Technology is transforming biomedicine from a predominantly descriptive science into a data intensive modeling science

EDELFELT, portrait of Louis Pasteur, 1885 The HUNT Biobank NTNU, Norway

Illumina X10

Or outperforming Moore’s law……

Craig Venter’s Longevity

The Baseline study by Verily

Digital data in the cloud. The Industrial Internet.

Consumer cloud solutions

Precision Medicine-Precision Health

Google trends Personalized Medicine vs Precision Medicine

Personalized Medicine became a buzzword after 2005

PLoS Medicine Vol 2, Issue

8, August 2005

Nature Vol 405

June 2000

WE NEED NEW STANDARDS

MATHEMATICAL MODELLING

AND SOLID STATISTICS

“All models are wrong but some are useful.”

p-values need to be backed up with biology

Pushing for reproducibility

The Synapse platform for collaborative science

Synapse by Sage Bionetworks

Synapse or GitHub guarantees provenance and reproducibility

Find an example at www.mispcamp.org

Check out TED talk style presentations available

under The Precision Medicine Revolution under

MEDIA at www.mispcamp.org

Literate Computing

Publish your research as virtual machine containers

http://www.nature.com/sdata/policies/repositories

“Open” Data Repositories

C++ for speed and power.

R and Python calls C++ when speed is needed

Galaxy is and Open Source data management platform

And there are several commercial solutions as well, like SevenBridges

Current global initiatives for data standardization and data exploitation need to

be broadened, deepened and propelled by a catalyzing ecosystem built by key

global partners dedicated to sustainably work together within a joint modeling

and analysis space supporting innovation

A coordinated network of resource hubs covering all continents and providing bio-banking, data generation (sequencing, spectrometry, cytometry, etc.), data storage and high performance computing resources federated through the OHIS

Global Health Genomics Center

A network of private non-profit and national institutes 33 Institutes/ 27 countries

A community of scientists, laboratory and public health experts

Acting locally as a Global Network

A backbone for the global ecosystem: The Institut Pasteur International Network- A global network of 33 Institutes

Aligning bio-banking, data generation, data storage, data management and data modeling, analysis and visualization.

Federating the network of resource hubs that provide infrastructure and resources

The collaborative platform: The Open Health Innovation Space

GHGC: Resource hubs for sample collection and biobanks

HUNT BIOBANK NTNU Norway

L. RUBENSTEIN/BROAD INSTITUTE BIOBANK

GHGC: Resource hubs generating standardized data

GHGC: Resource hubs performing reproducible data analyses leading to biological insights

The Healthy Human Global Project

Five countries, variable genetics, diverse lifestyles & environment

Establishing population-specific baseline measurements

Bringing diagnostic tools to the world Building on the Milieu Interieur Project, http://www.milieuinterieur.fr/en

Flagship projects powered by the Open Health Innovation Space. Example: The Healthy Human Global Project (HHG) Delineating the boundaries of a healthy human immune response Providing the base line for precision therapeutics

Whole blood

50ml

Nasal swabs

/Stool

X multiple stimulations

Field testing

Clinical

data Serology

Multiple donors x multiple continents age and sex stratification

1000 eCRF

≥ 300 var / p

60.000

Supernatant

Tubes

≈ 50 var / tube

≈ 250 var / p

20.000

RNA

profiles

≥ 600 var / tube

≥ 3000 var / d

12000 FCS files

≥ 500 var / p

3 Panels 2000 Genotypes

750K var / p

1000

Enterotypes 16S rRNA NGS

The HHG project : Building capacity & an unprecedented data warehouse

SAGE Bionetworks The Synapse platform

Stephen Friend MD. PhD. Apple

Chairman and Co-Founder of Sage Bionetworks

• CD3 + CD28

stimulation

after 22 h

• Null

stimulation

after 22 h

Group of 6 non

responders clustering

with the null stimulated

Gene expression separating 25

healthy subjects from 25

CD3+CD28 stimulated subjects

in a PCA plot

Discovery of non responders

Population stratification

for e.g. precision

vaccine development

Biomarker discovery based on the stratification

A t-test between responders and non responders

yield highly significant gene signatures