Development of a geneticsbased

1
Development of a genetics-based computer model
for development of new crops
C. Eduardo Vallejos Dept. Horticultural
Sciences Plant Molec. Cell. Biology
Carlos D. Messina Jim Jones Dept. Agricultural
and Biological Engineering
2
Food production systems
Land-based Agriculture
Greenhouse production
Advanced Life Support Systems
Vehicle Food System
Planetary Food System
3
Genetic Improvement in Plant Production

• Based on hybridization and selection in multiple
  environments
• Genotype x environment interactions cant always
  be extrapolated without risk
• Limited application for ideotype design for
  Advanced Life Support Systems

4
Objective
To develop a systems approach for ideotype design
based on previously characterized alleles at
selected loci. Ideotype design will be derived
from gene by environment interactions
simulations, scenario analysis and optimization
5
Roadmap

• Framework Concepts
• Soybean case study
• Bean case study
• Conclusions

6
Top-down approach

• Predictability of phenotypes
• Systems approach for modeling and simulation of
  plant growth and development

• Physiological and biophysical basis
• Honor mass energy balance
• Many processes remain largely functional
• Genetic variability paradoxically phenotypic
• Modular structure

7
Systems Approach to simulate plant growth and
development
8
Simulation of plant responses to temperature and
photoperiod
1.0
T01
T02
PPSEN
Tb
TM
CSDL
Temperature (C)
Day length (hs)
Model 1/d f(T) x f(DL) stagei d
photothermal days
9
Linking Crop Models and Genetics
Management
Linear Models QTL CART Bayesian Networks
Recombinant Inbred Lines Near-isogenic
Lines Mutants
CROP MODEL
Phenotype
Environment
O2, CO2, T, PAR, Rainfall, Pressure, DPV
10
From Phenotype to Genotype Ideotype design for
ALS
Management
CROP MODEL
Ideotype
ALS CONSTRAINTS
Loci, QTL, genes
Environment
11
Searching a Complex Space Simulated Annealing

• Efficient search through large spaces
• Discrete continuous variables
• Resistant to local optima

QTL Trait A
QTL Trait B
12
Implementation Genegro-Soybean
Two Planting dates
Linear Models f (Ei)
CROP MODEL
Phenology Pod Number
Gainesville, FL 2001-02

• Near-isogenic lines
• Seven E loci (Ei)

13
Predicting soybean yield and maturity from
variety trials
Variety Trials Phenotype
Genotype
Satt 496
Planting dates Row spacing Irrigation
Williams 82
Vinton 81
Omaha
Linford
Savoy
Yale
Nile
CROP MODEL
Yield Maturity
351 bp
330 bp
Illinois, 7 Locations 5 years 1995-99
14
Model evaluation
Time to Maturity
Yield
11
Simulated
R20.75
R20.54
Observed
15
Gene-based approach improve realism in ideotype
design for target environments
Short Photoperiod mid-season Drought
Long Photoperiod Terminal Drought
e1e2e3E4e5e7
E1E2E3E4e5e7
Probability of Excedeence
Simulated Yield (kg ha-1)
16
Lessons from working with Soybeans
Gene-based biophysical models can predict crop
growth and development and their interactions
with the environment
Coupled to SA can identify more productive crops
and increase the ideotype realism
Soybean as a model organism has limitations
arising from a complex genome and reduced genetic
variability for certain traits
17
Baseline Crops for ALS Program
Type of Research Focus
Vehicle Food System
Planetary Food System
Basic Research. Emphasize fundamental research
and feasibility studies.
cabbage, spinach, radish, onion
dry bean, rice
Concentrated Effort. Tech- nology development and
lab validation, including genetic selection,
yield optimization and cultivation management
carrot, chard, tomato
sweetpotato, peanut
Ready for Use in ALSSITB Limit research to
systems integration and model development.
lettuce
wheat, white potato, soybean
18
Bean as a model organism to enhance development
of Gene-based simulation models
Large Germ Plasm Collection 40,000 Accessions
Small Genome 632 Mbp
BAC Libraries 4.5X 10X fold
Linkage Map ca. gt800 molecular markers
Recombinant Inbred Family
19
Phaseolus vulgaris Linkage Map
20
Recombinant Inbred Family
x
f 1 (0.5)n-1
F1
F2
Fgt7
21
Analysis of Quantitative Trait Loci (QTLs)
Marker Locus
22
Where Do We Stand? Field Experiment, 2003, FL

• Biomass accumulation Partitioning
• Leaf Area
• Specific Leaf Weight
• Seed Pod number and mass
• Threshing (seed to pod mass ratio)
• Phenological Stages
• Growth habit

• Model evaluation and diagnostic
• Identification of QTL regulating important traits
• Data for deriving equations to predict model
  coefficients

23
Towards Genegro-Bean
Incorporates function of seven loci PPD, HR,
FIN, FD, SSZ-1, SSZ-2 SSZ-3 Approach Linear
Models CART
Yield (kg ha-1)
Seed weight (mg)
Simulated
Observed
24
A Glimpse into the future Thinking Globally and
acting locally

• Genegro-bean improvement using QTL and BN
  analysis
• Design ideotypes for ALS using SA and the
  Genegro-Bean
• Increasing understanding of biological process at
  molecular level will allow us to
• redesign or develop new MODULES that simulate
  gene network controls and metabolic processes
• use these to replace empirical parameters in crop
  models

25
Research integration via systems Approaches
26
Thanks!