Precision medicine require high precision in all parts which contributes to it. This presentation outlines the methodologies and pin point a subsystem whose precision is lot to be desired.

1. Personal Genomics, Population Centric Phenotypes
with Big Data – Fun for Statisticians
Athula Herath
Statistical Director/Translational Sciences
MedImmune, Cambridge, UK
12th
June 2015

2. Outline
A very brief introduction to hacking the genomes for
health
●
Good/OLD ways,
●
Contemporary ways,
●
Personal Genomes and beyond

3. Genetics and Genomes
(Good All Ways)
●
Explorations around a disease associated alleles
(unit of interaction) within the genome – one allele
at a time
●
Associate an allele to a “phenotype”/trait –
outcome(s) of interest for a malaise/malady
●
Sequence the allele in many individuals and
study the effects/outcomes carefully by
●
Cross-sectional, case-control, cohort and/or
randomized control studies (if we are lucky).
●
Falls very much within the traditional study
design paradigm

4. Genetics and Genomes
(Genome wide Association Studies)
● Then came the genome wide association studies (
GWAS)
● Explore Many/all relevant alleles (genetic
variants/genotypes) – primarily single nucleotide
polymorphisms – SNPs in one “disease” setting (or
a trait).
● Case-Control Studies (sequence all the
relevant/known SNPs)
● Some diseases (traits) are broad in scope some are
narrow (well defined/ill defined).

5. Personal Genomics
●
Consumer Genetics in the Social Network age
●
Whole populations irrespective of their health
(disease/trait) status – “normal” – (who is normal
anyway?) and “diseased” (and many will have many
diseases – we are all sick with many illnesses :( )
●
Made possible by
●
Costs of sequencing – cheaper than a cheap smart-phone >$100
($99 from 23andme – ~1000000 SNIPS)
●
30000, disease/health associated SNIPS identified by 23andme
●
>$1000 whole genome sequencing (Illumina – for large scale users)
●
Enthusiasm by the “networked” society.
●
Entrepreneurial start-ups (23andme, and now with the help from
HealthKit compilers and others)

6. Issues
●
First two approaches are fine, but they are either very, very, very
slow or very slow!
●
The third approach has the potential to be fast, but, as it is a way of
looking at many/unknown (phenotypes) to many (gigantic)
genotype interactions (just an under-estimation)
●
The accuracy of assaying/measuring the genotypes (sequencing
the genomes) has increased and of to an acceptable level/precision
●
The accuracy of the measurement of traits/phenotypes is falling
behind (some say at best it is hearsay :) – i.e. questionnaires)
●
So the precision of the assessment of the interaction between the
genotype and the phenotype may not be at an acceptable level
●
Tragedy: Astrology Reloaded – “My parents lives were haunted by“My parents lives were haunted by
'Horoscopes', if we don't do anything my children’s lives will haunted'Horoscopes', if we don't do anything my children’s lives will haunted
by 'Genomes' :)”by 'Genomes' :)”

7. Total Evidence Synthesis for defining
more precise phenotypes/traits
(A Challenge for all us/a rewarding occupation for some of us!)
Synthesize data that carry health
status/disease/traits (electronic
health records, and other Big Data
sources) to construct reliable
population centric phenotypes that
can be matched with this (other) Big
Data (genotype data)

8. 8
Landscape of Big Data in Healthcare
"..., big biomedical data are
scattered across institutions
and intentionally isolated to
protect patient privacy. Both
technical and social challenges
to linking these data must be
addressed before big
biomedical data can have their
full influence on health care.“
Finding the Missing Link for Big
Biomedical Data
Griffin M. Weber, et. al, JAMA.
2014;311(24):2479-2480.
doi:10.1001/jama.2014.4228
http://jama.jamanetwork.com/article.aspx?articleid=1883026

9. Sources of data for total evidence synthesis on
populations
http://jama.jamanetwork.com/article.aspx?articleid=1935130

10. The bigger point, he says, is that medicine will soon rely on a kind of global Internet-of-
DNA which doctors will be able to search. “Our bird’s eye view is that if I were to get lung
cancer in the future, doctors are going to sequence my genome and my tumour’s genome,
and then query them against a database of 50 million other genomes,” he says. “The result
will be ‘Hey, here’s the drug that will work best for you.’
http://www.technologyreview.com/news/532266/google-wants-to-store-your-genome/
Total Evidence Synthesis
… let me see :)

11. To some seeing is believing, but you
see what you look at – Makes you
think!
Enjoy! – Athula
(I hear Måns Zelmerlöw – “we are the heroes of our time” – http://goo.gl/2kaamm)
https://goo.gl/GVwgi3