Draft 1 Date 11/12/11 Page 1 of 9

raspgiantsneckΔιακομιστές

9 Δεκ 2013 (πριν από 3 χρόνια και 11 μήνες)

134 εμφανίσεις

Draft 1 Date
11
/1
2
/11


Page
1

of
9

Briefing paper: Bioinformatics projects and activities in King’s Health Partners


KHP has considerable expertise in Bioinformatics and Computational Biomedicine, spread over different
campuses. The Bioinformaticians located at the Computer Science Departme
nt, Strand Campus are mostly
focused on development of algorithms, with strength in Search algorithms, String Theory applied to pattern
recognition, Data Mining and Classification. At the IoP there is a strong focus in omics analysis including next
generat
ion sequencing, systems biology approaches, biomarker studies, data mining and classification. On
the Guy's Campus there are different groups focused a)
Genome Informatics
, mostly in the Genetics
Department and the cBRC; b) Cancer Bioinformatics at the Div
ision of Cancer Studies focused on
microarrays, miRNA studies and biomarkers; c) The Institute of Mathematical and Computational
Biomedicine, focused on statistical mathematical analysises, systems biomedicine and structural
bioinformatics.


This paper ai
ms to
(1) provide a vision for the future of bioinformatics at KHP and
(
2
)
describe

current
capabilities

that will form one part of a wider survey of informatics capability across KHP. Inevitably it will
remain a work in progress and we recognise that many

other KHP researchers are critically involved in the
projects described briefly below. Initiated by Richard Dobson and Franca Fraternali under instruction from
Mike Denis, lead for the KHP Informatics Grand Challenge, we welcome corrections, additions and

dissemination of the paper widely within King's Health Partners.


Section 1. Vision for the future:

Richard Dobson, Franca Fraternali, Michael Simpson, Thomas Schlitt


Kings Health Partners is impressive in terms of it’s current bioinformatics capabilit
ies, however it suffer
s

in
organization and coordination of initiatives, due its fragmented spread of resources and people. We have a
number of small groups each of which display considerable expertise and academic excellence.


The vision for future develo
pment is to unify and consolidate this expertise and wider health informatics
initiatives across KHP by creating a single presence. This presence
will

be steered
by

a coordinating
executive comittee, but
will require

centralised activities and resources. W
e
aim
to
establish
this presence in
a number of areas and coordinated research focuses. We currently fall short in service led bioinformatics,
training provision

and a central physical
presence
.


Below we lay out a number
key components for our vision

that

we feel will provide bioinformatics at KHP
with

a global presence
. This strategic plan is

in part inspired by models implemented by
leading

institutes in the
field of translational bioinformatics
including

Vanderbuilt University
and

the Dana Faber institu
te. These
institutes have developed model
s

that include a core of bioinformaticians, clinical informaticians, biomedical
informaticians and biostasticians
providing

analytical service support, administration, software development
whilst
integrated with aca
demic

research groups. Specifically, our focus for improvement is in four areas:


1
-
Physical Hub:

Moving forward, a k
ey

component for

bioinformatics within KHP
is
a unified

physical
presence, or hub. We propose an appropriately badged site that will house
training facilities, service support
and research space. We envisage a modern, fit
-
for
-
purpose site that would provide a world leading
environment for training and research and would aid in the recruitment of high calibre staff and students. The
hub will h
ouse
a core of non
-
academic, service led bioinformaticians and developers

and
also provide
opportunity

to house
existing

research groups
. Such collocation would create a critical mass of expertise

through which collaborations would be formed
, academic exce
llence maintained

and service delivered.


2
-
Computing infrastructure:

Ideally
the proposed

hub
will
house current computing infrastructure, presently
managed separately in different locations and administrered by isolated systems administrators. By
consoli
dating this infrastructure into a single data centre, there
will
be multiple benefits: fewer systems
administrators, with greater coverage; better use of storage and distribution of resource and greater
purchasing power.
There may
be logistical

challenges

including

requirements for
certain infrastructure
to be
housed at specific locations
, to access patient records for example
.
In such situations

we would
recommend

fast dedicated connections between infrastructure nodes and
the
physical hub where administra
tion
will
be

delivered. This resource
will
be delivered as autonomously as possible.


3
-
Training:

With the widely accepted challenge of the data deluge bearing down upon us,
we must ensure
that all staff and students passing through or resident within KHP
have the core informatics skills that are
increasingly essential for contemporary biomedical research.
We would
extend our current offerings and
develop
targeted
training programs

to fit within the current graduate and postgraduate educational
Draft 1 Date
11
/1
2
/11


Page
2

of
9

frameworks.
These programs
would provide

basic science and

clinicia
l students

with the
essential skills

required to design and perform statistically and computationally sound

analyses independently
.

In addition, a
dedicated PhD training programme specifically in bioin
formatics with co
-
supervision provided by clinical
researchers would provide huge opportunity for fostering new collaborations and would be an ideal platform
for establishing training and research excellence in critical areas. Such studentships located wit
hin the
proposed hub would be highly prized and would provide world
-
leading training in a world
-
leading
environment.


4
-
Leadership:

Within KHP there are a number of emerging leaders in informatics research, these individuals
are not only excelling in niche

areas but are opinion formers within the informatics field in general. We
therefore do not feel it necessary to recruit a single figurehead, but rather that current staff can implement
these proposals jointly and autonomously. Researchers will be far more

willing to contribute if they feel
empowered and we therefore feel the shared vision of the group would encourage participation from all
involved, more so than having to subscribe to a vision provided by a single figurehead. The most effective
way forward

would
be the creation of an executive committee who would oversee the consolidation of
bioinformatics within KHP (a unified hub), organise specialised teams in critical areas, recruit specialised
dedicated technical posts and lead the development of in h
ouse software that would be dedicated to the
needs of KHP and would allow for flexibility and adaptability.
We feel that s
uccessful

realisation of this

vision

would
make KHP very attractive to academics at all levels.



Section 1.

Current capabilities

Gro
up: Kathleen Steinh
ö
fel
, affiliation Computer Science Department, Strand Campus, KCL

CT Image Classification:

computer assisted radiology, and in particular, development of CT image
classification algorithms to support the diagnosis of focal lesions in li
ver tissue.

EPSRC project on Stochastic Local Search Algorithms for Structural Proteomics

(2006
-

2009) to
investigate prediction algorithms for protein folding and, in particular, local

minima of folded structures and
suboptimal

foldings of proteins with

applications to drug design.

Analysis of Mammographic Images:

development of algorithms for the analysis of mammographic images.
This research was supported by the Royal Society and Dr Sonia

Tangaro

visited

me

at King's for 4 weeks.

Personalised Medicine
and mircoRNA target Prediction
:

more
recent research aims at personalised
medicine, focusing on disease pathways induced by microRNA, is a new class of RNA that control many
celluar functions.

Key Contacts

Kathleen Steinh
ö
fel (
kathleen.steinhofel@kcl.ac.uk
)


Group: Sophia Tsoka
, affiliation Computer Science Department, Strand Campus, KCL

Analysis of biological network properties
: Biological entities organise in clusters/modules according to
functional and

evolutionary constraints. This work aims to develop algorithms to partition a network in
modules through optimisation methods. Algorithms have been developed for binary networks, weighted
graphs, as well as overlapping communities. Currently, extension of

these methods is underway for networks
that may change through time (dynamic networks).

Analysis of the effect of gut bacteria in rats:

Analysis of transcriptomics data to look into the affects of gut
bacteria. Microarray data from several tissues (e.g. l
iver, colon, ileum etc) are being analysed to find
differences between conventional and germ
-
free rats. Analyses will also be complemented by metabolomics
measurements.

Machine learning methods for data classification in biomedical data:
Given a microarray

experiment or
other types of high
-
throughput biochemical characterisation of patient samples, we develop strategies to
classify samples in appropriate phenotypes (e.g. disease type) according to observable features (e.g. gene
expression intensities). This

not only helps to characterise unknown samples, but can also serve to
distinguish which genes may be more important to a particular phenotype. Methods are based on decision
trees or hyperbox classification principles. These methods are applied on skin dis
ease data (psoriasis) and
breast cancer.

Draft 1 Date
11
/1
2
/11


Page
3

of
9

Analysis of expression in psoriasis skin disease data:

Expression data from a mouse model of psoriasis
analysed to reveal which pathways control the disease and its treatment with relevant cytokines.

Network analysi
s in cardiovascular medicine:
Relevant networks from physiological and pathological
cardiac hypertrophy have been constructed from biological experiments and analysed to derive their
topological properties. Comparative analyses showed significant differenc
es that were used to guide
proteomics experiments.

Key Contacts

Sophia Tsoka (
sophia.tsoka@kcl.ac.uk
)



Group: Schalkwyk, Leonard, SGDP, IoP


Research interests:

genomics, epigenetics, genetics, especially mou
se.


Computing

facilities:

132 processor Linux cluster with aggregate 376
Gb RAM and 29 Tb storage


Group:

Senior postdoc, system manager, PhD student in informatics part of program. We offer a yearly
course in the R statistical computing environment, thi
s year as part of the SGDP

summer school:
http://www.kcl.ac.uk/schools/summerschool/si/sgdp/


Key Contacts

Leo Schalkwyk
(
leonard.schalkwy
k@kcl.ac.uk
)


Group: Eric Blanc, BMS MRC Centre for Developmental Neurobiology, Guy's Campus


Research interests:

statistical methods to process and extract signal from large data sets, especially
expression micro
-
arrays. In particular, clustering methods

and co
-
expression network reconstruction are
among principal interests. Eric is also involved in the automation of the recording of climbing assays in flies.
Experienced in microarray data processing, programming (including R), and Bayesian statistics.


S
uggestions developments for

bioinformatics across KHP:

a) for meeting present bioinformatics needs in the college: central computing facilities with staff (technical for
system admin & scientific for bioinformatics support)

b) for improving research: PhD s
tudentship programme.


Key Contacts

Eric Blanc (
eric.blanc@kcl.ac.uk
)


Group: Institute for Mathematical and Molecular Biomedicine (IMMB), Guy's Campus


Led by:

ACC Coolen (Maths), F Fraternali (Randall Division)

The IMMB spearheads the College’s research and teaching activities at the interface between biology,
medicine, mathematics and computation. It aims to become a leading research centre devoted to the
development of quantitative tools for biomedical problem
s, and an efficient source of mathematical and
computational expertise for biomedical researchers. It will contribute to training a new generation of
biomedical researchers with a strong theoretical background, and organise workshops, conferences, and
shor
t courses.


The IMMB’s research is centred around mathematical and computational aspects of the following themes:

1.
Signalling and cooperation in complex intra
-
cellular and inter
-
cellular networks (PIs Fraternali and
Coolen)

current projects:

(a) Informat
ion
-
theoretic analysis of cellular signalling networks

Development of rigorous mathematical and computational tools for quantifying structure and complexity of

protein
-
protein interaction networks (PPIN) and gene regulation networks (GRN), with application
s in (i)
comparative interactomics, (ii) the generation of unbiased null model networks for hypothesis testing, (iii)
decontamination of PPIN and GRN data for method
-
specific experimental bias, and (iv) Bayes
-
optimal
Draft 1 Date
11
/1
2
/11


Page
4

of
9

detection of modularity. Collaboration
with NIMR (Kleinjung) and Universita di Roma La Sapienza (De
Martino). Funded by EPSRC and BBSRC.

(b) Proteomic reaction equations

Mathematical and numerical analysis and parameter exploration of the EGFR signalling system.
Development of nonequilibrium st
atistical mechanical methods for the analysis of complex formation and
dissociation dynamics in large protein interaction networks, based on generating functional analysis.
Collaboration with various partners at KCL, RAL, and the University of Oxford. Fund
ed by BBSRC (LOLA).

(c) Analysis of inter
-
cellular communication networks

Mathematical modeling of neural information processing systems (based on synaptic communication) and

of immune networks (based on communication via cytokines), including adaptation (
learning) and dynamical

response to perturbation (operation), and with a focus on understanding and modeling known pathologies

such as lymphocytosis and autoimmunity. Collaboration with Universita di Roma La Sapienza (Barra).

(d) Theory of epigenetic cell
reprogramming

Development of mathematical theory describing the reprogramming of gene regulation networks, by
intervention at the proteomic level, in order to achieve targeting cell phenotype modification. Based on
simplified but explicit description of th
e proteome
-
transcription
-
production
-
proteome cycle, and on theoretical
methods developed earlier (including programming protocols) for the manipulation of neural networks.


2.

Computational biomedicine Structural Bioinformatics (PI Fraternali)

current project
s:

(a)

Role of flexibility in molecular recognition The present project is exclusively computational: we hope to
characterize
in
-
silico

these dynamical properties of protein folds and to suggest new experimental work to
confirm our findings.
Funded by Leverhul
me FF PI.

(b)

Transcriptional programs in melanoma metastasis
.
We will focus on identifying gene patterns driven by
Rho GTPases and controlling the two types of movement. Microarray data analysis coupled to Protein
-
Protein interaction analysis will generate su
b
-
networks characteristic of the underlying phenotype.
Mechanistic studies with RNAi and over
-
expression approaches and in vivo studies to understand the
roles of these genes in metastasis will be carried out.

Funded by Cancer Research UK (CRUK) FFCo
-
appli
cant, collaboration with Vicky Moreno, Anne Ridley Tony NG

(c)

Validation of the APOBEC3G
-
Vif interaction as a drug target.
Funded by Wellcome Trust FF Co
-
applicant, collaboration with Hendrick Huthoff and Michael Malim.

(d)

Novel tools to map allosteric networks
in proteins. We aim to parameterise our method using a test
-
set of
allosteric protein structures, to apply the parameterised method to the Abl kinase of the oncogenic Bcr
-
Abl protein and to develop the software into a coherent package.
Funded by BBSRC FF P
I.

(e)


Protein
-
protein interaction networks and DNA methylation.
A long
-
term objective of the presented
research could be the prediction of methylation sites and the effects of environmental factors onto these
occurrences in healthy and diseased individuals.

The partnership with Nestle’
-
research centre in
Lausanne is particularly interesting because one of their main focus is on ‘Good Food, Good Life’ with
specialised research on nutrition and health.
Funded by BBSRC FF PI Partnership with Nestle’
Lausanne.


3
.
Stochastic methods in medicine

current projects:

(a) Bayesian analysis of FLIM/FRET imaging data

Development of mathematical and computational methods for the Bayes
-
optimal extraction of molecular
information from fluorescence imaging data in the low pho
ton number regime, to improve significantly the

timescales over which proteomic events can be detected in vivo, and to handle fluorescence lifetime
heterogeneity. Collaboration with the University of Oxford (Vojnovic). Funded by EPSRC.

(b) Survival analysi
s for heterogeneous cohorts with competing risks

Development of a Bayesian generalisation of survival analysis that allows for cohort heterogeneity and
informative population
-
level competing risks. The method leads to formulae for extracting regression
par
ameters and identifying latent classes, tools for quantifying heterogeneity and risk correlations, and
formulae for overall and individual survival curves that are decontaminated for the effects of heterogeneity
and informative competing risks. Collaborati
on with the University of Uppsala (Holmberg, Garmo). Funded
my Prostate Action.

Draft 1 Date
11
/1
2
/11


Page
5

of
9

(c) Dimension reduction and biomarker integration

Development of Bayesian models for dealing with fundamental problems in predicting medical outcome from

gene expression data,
viz. dimension mismatch (high dimensional signals with modest cohorts, leading to

overfitting and nonreproducibility of signatures) and integration with other biomarkers (clinical, imaging or

molecular). Dimension reduction is based either on using PPIN/GR
N data as constraints, or on unsupervised

latent variable (‘metagene’) analysis. Collaboration with KCL and international partners (FP7 network

IMAGINT) and Imperial College (Guo). Funded by EU and EPSRC.


Key contacts

Franca Fraternali (franca.fraternali@
googlemail.com), Ton Coolen (ton.coolen@kcl.ac.uk)


Group:Statistical Genentics Unit,
Division of Genetics and Molecular Medicine, Guys & IoP

Le
d by Cathryn Lewis and Mike Weale:

The Statistical Genetics Unit (SGDP, IoP, and Genetics & Molecular Medicine,
SoM) has substantial capacity
in analysis of genetic data (genotype and expression), integrating with clinical data, to identify and
characterise the genetic contributions to complex disease and traits. Four academics, 6 postdocs/fellows, 4
students. Capa
city to provide expertise in study design, analysis and interpretation. We have no core
funding for consultancy
-
level statistical analysis, but encourage longer
-
term collaborative research, joint
-
research projects and PhD students.
http://www.kcl.ac.uk/medicine/research/divisions/gmm/sections/clusters/statisticalgenetics.aspx

The four
academics are Cathryn

Lewis, Mike Weale
, Tom Price and Fruh
ling Rijsdijk.

Cathryn Lewis:

Research includes genetic risk estimation

-

statistical research programme with allied
software to determine an individual's risk of developing a disease using genotype data, combined with
relevant epidemiological risk facto
rs and clinical covariates. Capacity to perform population
-
wide analyses to
determine risk profiles (e.g. what proportion of the population will be at >5
-
fold increased risk of disease),
and individual
-
level risk estimation. Risk categorisation based on
confidence with which information is
known. Current projects in schizophrenia, rheumatoid arthritis, Crohn's disease, with three clinical research
fellows.

Mike Weale:

Research

component focuses, in part, on the integration of different types of data in
order to
make better joint inferences about, for example, whether or how a particular gene is causally affecting a
particular medical trait. Particular projects include: (1) the statistical integration of genomic annotations with
genome
-
wide association s
ignals; (2) the joint analysis of genotype with gene expression data in different
brain tissues; (3) methods for detecting gene
-
gene interaction signals in large genetic datasets; (4) methods
for inferring the origin of unknown forensic DNA samples; (5) st
udy design and power calculations for Next
Generation Sequencing experiments


Facilities:

In terms of facilities and equipment, regularly use but do not own two computer clusters based at Guy's
Campus (Athena and Hera) and one based at SGDP (Mumak

belongi
ng to Leo Schalkwyk
).


Key contacts

Cathryn Lewis (
cathryn.Lewis@genetics.kcl.ac.uk
), Mike Weale
(mike.weale@genetics.kcl.ac.uk)


Group: Rebecca Oakey,
Division of Genetics and Molecular Medicine, G
uy's campus

Skills, areas/interests: G
roup focuses on the epigenome and understanding the methylome, histone
modifications and the role of DNA binding proteins in health and disease using next generation s
equencing
and bioinformatics.

F
acilities/equipment
: Equipment

is housed and operated by the BRC and KCL Genomics core facilities
located on the 7th floor of the Tower Wing on the Guy's campus and comprise: three

Illumina Genome
Analyser IIx (GAIIx) sequencers, one HiSeq 2000, two cBots,Applied Biosystems

7900HT Real
-
time PCR
system, Covaris Adaptive Focused Acoustics The computing facilities are also owned and managed by

t
he
BRC:

Draft 1 Date
11
/1
2
/11


Page
6

of
9

HPC Cluster
(to which Mike Weale and Cathryn Lewis of the SGU refer)
comprises some 33 servers for
computation provided by 16x

IBM iDataplex 2U chassis containing 32x IBM IDataplex nodes each fitted with:



2x 2.66Ghz Intel Nehalem processors (8 cores per node and a total of 256 processor cores for the
cluster excluding heads nodes and the large node)



48GB memory



250Gb hard driv
e

A larger compute node an IBM x3755 server fitted with:



2x 2.6Ghz 6 core processors (a total of 12 cores)



128GB memory



2x 146Gb hard drives

A Small HPC Cluster comprises some 9 servers for computation with:



4x HP G6 x86_64 compute nodes (2.6GHZ, 28G
B RAM) running



4 Tesla S1070 Nvidia GPUs for novel GPU processing applications



3x Supermicro compute nodes (2.6GHz, 32GB RAM)

S
ize and details of the group
:


R
esearch group has fiv
e basic scientist researchers. C
ollaborate with RCUK fellow Reiner Schulz

and his
PhD student, both bioinformaticians and work together with the
s
BRC Systems administrator and two
bioinformaticians.

Key contacts

Rebecca Oakey (rebecca.oakey@genetics.kcl.ac.uk)


Group: Nicolas Smith
, Head of Biomedical Engineering, Imaging Scien
ces & Biomedical Engineerin
g
Division, St Thomas’ Hospital

C
urrent interests are developing multi
-
scale models from gene regulation to whole organ function which
integrate experimental data at each level to provide a mechanistic framework for analysing phy
siological
fun
ction. The majority of

activities are in the cardiovascular systems although we also work in cancer and in
both cases there is a significant focus on the clinical translation of these models.


S
kills are in computational simulation of physiol
ogical systems, development of efficient numerical
techniques, analysis of sub
-
cellular and cellular regulation and the integration of cellu
lar and whole organ
responses.

Facilities:
Central to these facilities is the 500 processor shared memory High Perf
ormance computer which
is

about to

be acquird

and a range of imaging scanners. The department of biomedical engineering is 15
academics with 7 specifically focused on computational modelling.


Key contacts

Nicolas Smith (nicolas.smith@kcl.ac.uk)


Group:
Ri
chard Dobson
,
Lecturer, Bioinformatics lead,
BRC
-
MH, South London and The Maudsley &
IoP

The assembled team
of 6 postdocs
have developed expertise in the analysis, integration and modeling of
complex large molecular datasets such as expression and SNP arra
ys, next generation sequencing,
approaches for network (gene regulatory, co
-
expression and protein in
teraction) and pathway studies

providing novel insight into disease mechanisms and biomarker discovery
.

We

are performing a range of studies through the co
mbined analysis of clinical, imaging, proteomics,
transcriptomics and genom
ic
datasets
, and have
multiple
active academic
and industrial
collaborations
.

Draft 1 Date
11
/1
2
/11


Page
7

of
9

We

are developing infrastructure to enable sharing of datasets supported by funding for a pilot project

from a
joint KHP/UCSF initiative in collaboration with the NIH funded Clinical and Translational Science Initiative
(CTSI).

Through grant funding
we

have expanded the high performance computing Linux cluster
(http://compbio.brc.iop.kcl.ac.uk/cluster/index
.php), purchased with NIHR capital finding (August 2009) to sit
within the SLaM firewall. This now enables them to text mine all SLaM patient records (>170k) in under 1 day
and align a lane of next generation sequencing against a human reference genome in
hours rather than
days. This has significantly enhanced capacity to support the rapid analysis and interpretation of large
variable imaging and 'omics datasets.
Utilising this infrastructure the imaging theme has introduced
automated quantitative neuroimag
ing into the routine assessment of MRI scans for people with dementia.


Key contacts

Richard Dobson (
richard.j.dobson@kcl.ac.uk
)



Group:
Matt Arno, Waterloo campus

Service led model. Includes: design of pr
oject, generation of data, data analysis with researcher, pa
thway
analysis

Key contacts

Matt Arno (matthew.arno@kcl.ac.uk)


Group:
Andrew Pickles
,
Department of Biostatistics & BRC
-
MH, IoP

P
rimarily a group of statisticians several of whom have expertise i
n bioinformatics methods and applications,
primarily in the area of expression and SNP array data and in neurophysiology, ERP in particular. We also
have expertise in formal statistical modelling e.g. of biomarker mediation models that account for
measurem
ent error, particularly in software such as Mplus and gllamm (a powerful and very general
modelling software that originated from this Department, see
www.gllamm.org
).


Key Contacts

Andrew Pickles (
andrew.pickles@kcl.ac.uk
)


Group:
Thomas Schlitt,

Lecturer in Bioinformatics King's College London Dept of Medical and Molecular
Genetics 8th floor Tower Wing Guy's Hospital


S
kills, areas/interests
:

The main resea
rch focus of the group is in the analysis of gene and protein networks.
We develop novel approaches for de
-
novo pathway discovery to support the discovery of genes and
pathways underlying complex diseases independent of pathway annotation. We have experien
ce in the
integrated analysis of GWAS/gene expression/NGS data and gene and protein networks. We are involved in
several next
-
generation sequence analysis projects such as analysis of exome data to find disease genes
Crohn’s disease. In collaboration with
Prof Ahlers (Hannover, Germany) we develop our network analysis
and visualisation tool BioGranat. And in collaboration with Dr Brazma (EBI) we work on dynamic models for
the simulatio
n of gene regulatory networks.


F
acilities/equipment
:

The group owns two

dedicated computer servers, one mainly used as MySQL database
server, the other is mainly used for computational tasks. We also make intensive use of the cBRC Athena
HPC cluster at the Department of Medical and Molecular Genetics. We maintain our own data
base of
protein
-
protein interactions which we integra
ted from six public databases.


S
ize and details of the group
:

Dr Thomas Schlitt (group leader), Nick Dand (PhD student 2011
-
2014), Russel
Sutherland (Illumina CASE PhD student, 2011
-
2015)


Key Contact
s

Thomas Schlitt (
thomas.schlitt@kcl.ac.uk
)



Group:
Emanuele de Rinaldis,

Senior Research Fellow, Breakthrough Breast Cancer Unit, King’s College
London, UK

Skills, areas/interests:

Integration of
-
omics

d
ata (SNP, CGH, miRNA, ExonArray, Methylation) and
clinical/pathological data for the study of the molecular mechanisms underpinning breast cancer

Draft 1 Date
11
/1
2
/11


Page
8

of
9

Discovery of novel biomarkers, prognostic factors and therapeutic targets in breast cancer, using genomics
tec
hnologies and bioinformatic/biostatistical data analysis.

Clinical
-
Genomics data integration and analysis: development of databases, tools and algorithms

Interest in NGS (although very new in the field)

Facilities/equipment:

as part of the Breakthrough Bre
ast Cancer Research Unit we have access to lab
facilities and work in close link with lab scientists for follow
-
up and validation experiments

-

3 personal workstations, 1 64
-
bit CentOS, quad core, 16GB DDR2 Linux Server, 1 High Performance
server: HP DL G5
380 Quad core X 2 2.83 GHz, 1 Storage server: HP MSA 2312, 6TB, 2 NAS Storage
Devices 4TB

Size and details of the group
: (1) Emanuele de Rinaldis (Biology/Bioinformatics)
-

Team Leader, (2) Anita
Grigoriadis (Biology/Bioinformatics)
-

Research Fellow, (3)
Brian Burford (Computer Scientist)
-

Research
Associate, (4) Akram Shalaby (Mathematician)
-

PhD student (co
-
supervisor: Ton Coolen)

How can existing capability be better coordinated and utilised and /or enhanced through investment and
development

-

We thi
nk it would be good to have a King's school of bioinformatics with attached training initiatives, PhD
programs, lectures, workshops etc. Maybe it exists already, if so we are not aware of it (in thsi case probably
we have a problem of communication to inve
st on). A nominated head of such a school would help
coordinating school's activities.

-

More in general, there should be more investments in training and exchange opportunities for young
bioinformaticians

-

Facilities for data storage, backup and administ
ration

-

Facilities for controlled data sharing across different bioinformatics King's groups

-

Dana
-
Faber Cancer Institute in Boston it's a good example of a place where bioinformatics has very strong
impact in cancer research

Key Contacts


Emanuele de R
inaldis (emanuele.de_rinaldis@kcl.ac.uk)

Group:
Michael Simpson,

Lecturer in Medical and Molecular Medicine, King's College London Dept of
Medical and Molecular Genetics 8th floor Tower Wing Guy's Hospital



The main focus of the group is the application
of contemporary genomic technologies to the elucidation of
the genetic basis of human diseases and traits. An integral

component of our research programme is centred
on genome informatics. We have established data analysis pipelines for the interpretation
of next
-
generation
sequencing data and continue to evaluate and develop novel methodologies for the analysis and integration
of large
-
scale genomic datasets to build a comprehensive understanding of the role of the genome and its
variation in health and di
sease.



Our d
ata analysis is principally undertaken on the GSTT BRC
HPC Computational Cluster.
We work closely
with the BRC Genomics and Bioinformatics Core Facilities and collaborate with both clinical researchers and
computer scientists within KHP, the
UK and beyond.



Key
Contact
s
:

Michael Simpson

(
michael.simpson@kcl.ac.uk
)


Group: Prof Michael Luck,

Dept of Computer Science, KCL

Members of the Group worked on various

topics on medical oriented research
, e.g. arrhythmia classification
on mobile phones, herpes virus identification. In particular they developed algorithms for mapping high
throughput sequencing technologies as well as weighted and degenerate sequences. The team also
developed the first tran
scriptome map of mouse isochores from three distinct mouse tissues (muscle, liver,
Draft 1 Date
11
/1
2
/11


Page
9

of
9

and brain).


The team developed NGS prototype software




REAL (Read Aligner): an efficient, sensitive, and accurate alignment programme that can match or even
outperform mo
st well
-
known tools, e.g. SOAP2, Bowtie, BWA




cREAL (circular REAL): an extension of REAL specifically designed for small genomes with circular
structure, e.g. bacterial chromosomes




DynMap: yet another alignment programme specifically designed for ali
gning a set of reads against many
closely related genomes, e.g. individuals of the same species


Members of the group worked on Haplotype Classification Algorithms, on disease gene identification related
to haematopoiesis,

Mammographic image analysis as w
ell as Combinatorial Algorithms for Protein Folding
Simulations and lattice protein folding simulations.


Furthermore team members worked on analysis of biological network properties, analysis of the effect of gut
bacteria in rats, machine learning methods

for data classification in biomedical data, analysis of expression
in psoriasis skin disease data, as well as network analysis in cardiovascular medicine:


Fundamental research topics of the group include:


● Combinatorics on words and graphs

● Probabilis
tic analysis

● Constraint programming

● Combinatorial optimisation

● Automata theory

Sequence analysis (alignment, re
-
alignment, etc.)

● Comparative genomics (evolution, function prediction)

● Classification of biomedical data

● Music analysis (rhythm, mel
ody detection)

● Structure and search of large networks


(www, p2p, wireless, etc.)

● Text compression

● Data mining


Key
Contact:


Michael Luck (
michael.luck@kcl.ac.uk
)


This paper will be updated on request.

Further information relating to health informatics in KHP should be
forwarded to:
richard.j.dobson@kcl.ac.uk

or

franca.fraternali@kcl.ac.uk