Transcript Analysis of Crop Plant Genomes

Analysis of crop plant genomes
Jo Dicks
John Innes Centre
[email protected]
http://jic-bioinfo.bbsrc.ac.uk/bioinformatics-research/
Data
We want to compare the genomes of
crop plants (e.g. wheat, rice, maize,
millets, barley, pea)
At present, we mainly compare:
Whole genome sequences
Genetic markers (comparative
mapping)
Transposable elements
What can we learn from the data?
Understand evolutionary processes in crop
plants.
Use comparative mapping to predict
gene/marker location and function across
species.
Use transposable elements to maximise
diversity within a subset of a germplasm
collection (core collection).
Whole genome sequences
Linear streams of data, where each
element is represented of one of four
letters (A, C, G or T).
Streams can be long – billions of letters.
Blocks of sequence can be meaningful
(e.g. they encode genes or transposable
elements) or are deemed ‘junk’.
Species 1: caggaaaacacacactcacatacatgaacaatatctc
||||| || |||||
|||||||| |||| || ||
Species 2: caggataatgcacac
catacatgcacaaaat tc
Comparative mapping data
Markers have a location and an orientation.
When markers in two species are related by
descent from a common ancestor, they are called
homologues.
Comparative mapping data are combinatorial.
Species 1
1
5
3
4
2
Species 2
1
2
4
5
In most data sets, links (homologies) may be spread across chromosomes
Retrotransposons
Retrotransposons are a type of transposable
element.
There are various locations in a genome where
they are either present or absent.
An entry in a germplasm collection (called an
accession) is therefore essentially a barcode
representing multiple retrotransposon locations.
Accession 1
1
2
Accession 2
1
2
3
4
4
Evolution
Data change in time due to errors known
as mutations (there are several distinct
types of mutation).
Differences between species are often
quantified in terms of the number and
type of such mutations.
The relationship between species is often
represented as a tree of evolution
(often called a phylogenetic tree).
An evolutionary tree
Ancestral species
Species 1
Species 2
Species 3
Species 4
Mutations occur through time, along the tree branches
Data problems
In comparative mapping studies, there
may be elements between the markers
that are important but of which we
know nothing (i.e. missing data) and
erroneous links between data items (i.e.
data errors).
Missing data will be largely alleviated by
whole genome sequences (when will
this be though?) but there will still be
errors in the data.
Projects
UK CropNet (data)
CHROMTREE (analysis)
GENE-MINE (data)
Germinate (analysis)
JIC are also involved in Arabidopsis
and Brassica IGF projects
UK CropNet databases
UK CropNet curates and develops
databases and data analysis tools for:
Arabidopsis thaliana (AGR)
Brassicas (BrassicaDB)
Cereals (BarleyDB, CeResDB and MilletGenes)
Forage grasses (FoggDB)
Potato (SpudBase)
as well as developing a database for:
Comparative mapping data (CropSeqDB and
ComapDB)
Problems
To get hold of comparative mapping data
from the crop plant community, we need
to access disparate data sources of
differing quality (not necessarily
electronic).
We need to link the data sources to form
a single, queriable entity.
The UK CropNet single- and related-species databases
AGR
BarleyDB
SpudBase
ARCADE
MilletGenes
BrassicaDB
CerealsDB
FoggDB
ComapDB
Will the GRID be a better solution than ARCADE?
Analysing chromosomal evolution
Chromosomes evolve over time
Inversion
Inversion
Translocation
Inversion
Mutations events can be mathematically modelled and
used to construct a phylogenetic tree
Problems
Unlike
DNA
sequences,
data
are
combinatorial, not linear.
Algorithms are very slow (many require
optimisation over a multi-dimensional space)
and analysis of large data sets is not
currently possible on JIC machines.
Parallelisation of algorithms may help, as it
has done for DNA sequence phylogenetic
analysis. However, is the only answer?
In some cases (due to mutations such as
allo-polyploidy) we may wish to consider
phylogenetic networks instead of trees –
an even harder computational problem.
GENE-MINE and GERMINATE
Analysing germplasm collections
Germplasm projects
GENE-MINE: An EU-funded project to
develop a data-management and analysis
computer system for plant germplasm
collections
GERMINATE: A BBSRC-funded project allied
to GENE-MINE and another EU project
TEGERM, to develop specialist tools for
analysis of the TEGERM data.
The problems seen in these projects are
essentially the same as those of UK CropNet
and CHROMTREE.
Retrotransposon insertion
1
2
3
Like chromosomal mutations, retrotransposon
insertion can be mathematically modelled
Relationship between accessions
INS
INS
INS
Again, sometimes we may need to estimate a phylogenetic
network (due to introgression between accessions)