ANS/PLSS 433: Molecular Breeding 2

III.	Construction of Genetic maps using RFLP inheritance Patterns.

Introduction:

	RFLP markers are inherited in a mendelian fashion. That map 
distance is a function of recombination frequency. That we can generate 
RFLPs easily from any chromosomal DNA fragment cloned in a plasmid. Today 
I'll describe the steps necessary to construct an RFLP map of the genome.

1. Select and Screen Parent plants

	To construct a map we need parents which are genetically divergent 
at many loci and in which desirable agronomic traits are segregating. Even 
if a map of the genome already exists for each new pair of parents in which 
we would like to map agronomic traits we must first detect useful 
polymorphisms. DNA is extracted from the parents and digested with a 
restriction enzyme, electrophoresed through agarose, Southern blotted and 
hybridized to one clone. RFLPs between the parents are noted and that 
probe/enzyme combination used again later on the F2 population. There are 
over 600 different restriction enzymes but we usually only screen 5-10. 
Polymorphism frequency between parents ranges from 80% per enzyme between 
crossable species to 20% per enzyme between distinct varieties to 0-1% 
between backcrossed varieties. To find 100 useful probes we might need to 
test 500 with 1-10 restriction enzymes.	Screening many varieties reveal 
some probes detect RFLPs 5 fold more frequently than others.

2. Produce a Mapping Population

	The selected parents (P1 and P2) are crossed to produce the hybrid 
F1. The F1 plants are selfed to produce the F2 or backcrossed to one parent 
to produce the B1. F2 populations give more information per sample. 
If scoring agronomic traits over several environments and several years we 
can use later generations of selfed progeny F3-F9. Often 50 individuals 
are mapped which gives maximum resolution for map distances of + or - 2cM, 
100 individuals give map distances + or - 1 cM and so forth.

3. Score RFLPs in the mapping Population

	DNA is isolated from individual plants and digested with 
appropriate restriction enzymes. After electrophoresis, survey filters are 
prepared for each enzyme used. Filters are hybridized with a probe, exposed 
to film and then scored. Filters are rehybridized 10-20 times with other 
probes.

4. Linkage Analysis

	Data is accumulated for 100 or more probes. Linkage is calculated 
by the degree to which probes tend to cosegregate. The number of linkage 
groups will correspond to the number of chromosomes. A typical genome 
contains 1300 cM, 100 probes will give one RFLP marker every 10 to 20 cM. 
As linkage distance increases the recombination frequency tends to 50% 
because double crossovers become as frequent as single cross overs. 
If markers are unlinked recombination frequency is also 50%. A map is the 
sum of many small linkage associations.

IV.	Breeding Applications of Gene Mapping

Introduction: 

	We have seen that using molecular probes we can detect polymorphism 
between alleles of a gene which allow us to map the plant genome. We also 
found that the same approach reveals the genotypes of the F2 generation. 
From this data we can select which individual will contain a useful 
combination of genes.

1. Indirect Selection: 

	Selection for a conventional gene might be expensive difficult 
and/or time-consuming (eg SDS resistance). Using RFLPs flanking a gene 
affecting resistance we can indirectly select for the RFLP instead of 
directly for the gene. If the gene is flanked on either side by RFLPs 
10 cM away then 99% (100% -[10% x 10%]) of the individuals with the RFLP 
markers would also carry the gene. This would be particularly useful for 
recessive genes. Normally progeny testing is necessary between each cycle 
of selection to ensure heterozygosity. Gene pyramiding would also be aided, 
this is where unlinked genes contribute to the same character. Progeny 
testing is usually necessary but RFLP markers would obviate this need.

2. Quantitative traits: 

	Most trait of agronomic significance are the product of a pyramid 
of genes spread across the genome. Each gene has a small effect on the 
phenotype which can be positive or negative! Co-dominance is often 
observed and environmental effects may be observed. Mendelian analysis 
breaks down. RFLP maps allows us to identify how many genes are 
contributing to a character and their relative contribution 
(what % of the variance observed) using simple statistics on means of P1, 
P2 and F1 class data (T tests, ANOVA, Interval Analysis). Then scoring 
QTLs by their linked RFLPs becomes simple mendelian genetics. However for 
each QTL to be scored a new cross and RFLP map needs to be constructed! 
We need a shortcut!

3. PCR based markers:  

	Markers rely on PCR and gel electrophoresis so take 1/7 of the 
time of RFLPs and 1/100 of the DNA. PCR is simply the enzymatic 
amplification of a DNA fragment flanked by primers of known sequence. 
We test primers to find some that reveal differences between parent. Then 
score them just as RFLPs (ARFLPs).
 
	a. Amplified Fragment Length Polymorphism (AFLP) are RFLPs of which 
		a subset are amplifed at once, by ligating linkers and the 
		amplifying subsets of the ligated molecules. They give 
		100 bands per lane, 10% polymorphic.

	b. As primers get shorter the PCR products lose their gene 
		specificity and codominance. These markers Random 
		Amplified Polymorphic DNA (RAPDs) and DNA Amplification 
		Fingerprints (DAF). These produce multiple bands like 
		barcodes. Each band behaves as a simple dominant mendelian 
		marker and so can be mapped. They give 10-50 bands per lane, 
		5% polymorphic. 


	c. Minisatellites or Simple Sequence Repeats (SSRs) are simple 
		dinucleotide repeats that frequently vary in length between 
		alleles of a gene, they are more effective than RFLPs for 
		allele identification. 1 band per lane, 50% polymorphic. 

	d. SSRs can be combined with AFLPs for multiple band analysis by 
		SAMPL (Selective Amplification of Microsatellite 
		Polymorphic Loci). 100 bands per lane, 50% polymorphic.

	e, Specific alleles of a marker can be tagged if the sequence is 
		known with SCARs (Sequence Characterized Amplified Regions).
		 1 band per lane, 100% polymorphic.

These technologies are amenable to automation and may help shorten the RFLP 
map for every QTL of interest.  

4. The importance of crop breeding: 

	Grain yields have increased about 2 % a year since 1930. 
Genetic improvement is responsible for 50% of this improvement, Fertilizer 
use increase 20% and improved Pesticides 25%. Relatively Crop breeding is 
the most important agricultural science. Markers very lose to important 
genes can be used to isolate the genes, for example the rps2 gene and pto 
genes were isolated this year. They are the first plant disease resistance 
genes cloned and hold promise for direct genetic manipulation of disease 
resistance in the future since they direct the hypersensitive response, a 
plant response common in all resistant genotypes of all species.
However, improvement in important crop genetic traits is often complex and 
polygenic. The genetics and physiology of improvement are little understood.
For the future we need to improve yield potential, stress tolerance and 
broad pest resistance. All normally polygenic traits. Transgenic approaches 
will make small temporary contributions to crop improvement but RFLP 
analysis of QTLs promises far greater and sustainable benefits.