Quantitative interpretation of gene network controls and the effects of selection on these

fancyfantasicΤεχνίτη Νοημοσύνη και Ρομποτική

7 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

81 εμφανίσεις

Project proposal idea for
CSS August
200
2

Quantitative interpretation of gene network controls and the
eff
ects of selection on these

Dr. Scott Chapman, Plant Industry,
St. Lucia

Prof.
G.L. Hammer
, The University of Queensland


A two year post
-
doc project


Two strategic
research
areas for
Australian science
are:

-

Genome/phenome research to address the need for biologists who understand molecular
genetics
and

the behaviour of the whole cell, the whole tissue and the whole organism.

-

Complex system control,
e.g.

as exhibited in plant phenological response to the environment

For CSIRO to maintain a lead in genome/phenome research, it must begin to develop skills in the
analysis and interpretation of complex gene networks and in the ways in which these can be
manip
ulated to improve crop adaptation. The proposed research would focus on physiological traits
that CSIRO PI works on and link to the current UQ/CSIRO/QDPI project

with Pioneer Hi
-
Bred (the
world’s largest plant breeding company)
. CIRAD researchers are also
keen to collaborate in this
area
, and a grant has already been obtained to visit them in France this winter
. During development
the plan would link with Canberra researchers in plant development and genetics

(Programs P and
X)
, including rice gene machine
team. The generic nature of the research could extend to
quantifying the control of other processes such as disease response. Some basic models exist for
these, but there are substantial development needs.

Molecular biologists, including in CSIRO PI are un
ravelling the genetic controls associated with
two of the most important genetic controls of crop yield:
flowering
time
and plant height. The
approximate genetic networks controlling these traits are frequently described in a qualitative sense,
using techn
ologies like mutation and gene silencing to knock out functional genes from a single
genotype. To be applied in the understanding of the process of plant breeding, these genetic
networks need to be (i) described and quantified to better infer their physiol
ogical effects; (ii)
investigated to determine the allelic effects that contribute to genetic variation. In conjunction with
CSE, CLW, QDPI, UQ and other private industry partners, we have and continue to develop two
types of models of genome/phenome attri
butes:

-

a biophysical model (APSIM) to propagate gene signals over scales (gene to canopy)

-

a gene simulation model (QU
-
Gene) to determine the effects of recombining genes that
control different biophysical attributes.

We use a cluster of >80 computers to ge
nerate the phenotypic state
-
space resulting from genetic
networks that differ in structure and strength of allelic effects. Our research would (i) help infer
genetic control from expression data; (ii) reduce the state
-
space according to improved
quantifica
tion of the genetic controls of flowering and height; (ii) improve the identification of
better breeding methods that exploit understanding of established and selectable gene networks.

The project will employ a post
-
doc to
implement
a biophysical simulatio
n module of
the gene
networks controlling
flowering and plant height interactions using data from the literature, small
CEF experiments & PI labs on
Arabidopsis
, rice & wheat. The model will be parameterised to
reflect known and postulated allelic variatio
n for various control points (gene networks) of these
processes & tested to examine which types of genetic networks appear to best reflect real
-
world
observations. Given the allelic variation, successful plant breeding strategies will be identified via
lar
ge
-
scale simulations where the biophysical and gene simulation models are linked.

The CSS outcomes from this work include:

-

data and computing methods to generate complex growth and development behaviour for
plant models

-

insight into the control of signalli
ng networks for key traits in crop plants

-

investigation of Bayesian networks in the prediction of plant phenotype from marker
information

Chapman, S.C., Cooper, M.
and Hammer, G.L.

(2003
).

Evaluating plant breeding strategies by
simulating gene action and dryland environment effects
.

Agronomy Journal
,
95: 99
-
113
.

Cooper, M., Chapman, S.C., Podlich, D.W., Hammer, G.L. (2002)
.

The GP problem: Quantifying
gene
-
to
-
phenotype relationships. In Silico Biology 2, 0013 On
-
line invited paper:
http://www.bioinfo.de/isb/2002/02/0013/

Resources

The methods are largely to be developed from known research data
.

The research would require
travel and accommodation funds to work with researchers in Canberra (one to two visits per year)
to
utilise past and present experiments on flowering and plant height.

At current
costs
,
the project
would require ca. $130
-
160 000
per year
for
2 years
.

The investment would greatly leverage current
relationships with a large breeding company and other in
stitutions to deliver insights into the
application of CSS in controlled plant evolution, i.e. plant breeding.