[To contribute to this conference, send your message to biotech-room2@mailserv.fao.org. For further information on the Electronic Forum on Biotechnology in Food and Agriculture see Forum website.
Note, participants are assumed to be speaking on their own behalf, unless they state otherwise.]

-----Original Message-----
From: Biotech-Mod2
Sent: 26 November 2003 10:12
To: 'biotech-room2@mailserv.fao.org'
Subject: 42: Re: When the marker is the gene

From Prof. H. Dulieu, again. I worked in the Plant Breeding Station INRA-Dijon until 1999. I was professor of plant biology and genetics at the University of Burgundy. I am presently retired.

Several participants - Adison Mota (Message 36, November 24), Hayde Galvez (Message 38, November 25), Don Nicol (message 39, November 25) - discussed the case of complete linkage (0 cM!) between the marker and the gene (trait) to be selected. Of course, if selection is made on the marker which has its allelic form completely associated with the allelic form of the trait, the probability of selecting for the trait (gene) rises to a maximum. However, it must not be neglected that a complete linkage group (one chromosome) is associated with the trait, then you perform a simple Mendelian experiment!

In many examples of crops or animals, the variety or breed to be improved has been previously selected and maintained, due to its high productivity or performance. A back-cross selection must then be performed over several generations to incorporate the trait(s) of a "donor" genotype into the existing variety which must recover its qualities. This implies selection over several generations, partially depending on the number of linkage groups (chromosomes). The genes located on the same chromosome as the trait must be recovered from the recurrent genotype (the receiver of the trait) by recombination. This is the reason why I considered (message 23, November 20) it important to have a map available with many molecular markers over all linkage groups. These markers are to be used also for counterselecting the regions of the donor genotype which must be replaced by the resident genes from the existing variety.

Theoretically, the ideal situation is to select the trait indirectly by molecular markers at a short distance (say 5 cM), to the left and right of the trait locus. The "false positives" should then be very reduced in expected frequency (p1*p2= 0.0025) because they must result from double crossing-over, where the trait locus ends up between the allelic forms of the markers of the receiving line. Then, for these reasons, I think that selection based on a single marker, even strictly associated with the trait, is not efficient as compared to phenotypic selection on the trait itself, especially if the breeder does not help the return to the existing variety with MAS applied to other markers located on different chromosomes, including the chromosome where the trait has been mapped!

Prof. H.Dulieu,
Plant genetics and Biology,
University of Burgundy,
France
hdulieu (at) u_bourgogne.fr

[Having a marker on either side of the gene of interest (Q) provides greater precision than having a single marker when following inheritance of Q from one generation to another. Consider a situation where a parent has the marker allele M1 and the allele Q1 on the same chromosome and the rate of recombination (crossing over) between M and Q is r (e.g. 0.05). If an offspring receives the allele M1 from the parent, there is a 0.95 probability it has also received Q1. With flanking markers, consider now that there is an additional marker N on the other side of Q, such that Q is midway between M and N and r is 0.05 between M and Q and between Q and N. Now, if a parent has the alleles M1, Q1 and N1 on the same chromosome, if an offspring receives the marker alleles M1 and N1 from the parent, there is a 0.997 probability it has also received Q1...Moderator].

-----Original Message-----
From: Biotech-Mod2
Sent: 26 November 2003 10:36
To: 'biotech-room2@mailserv.fao.org'
Subject: 43: Re: MAS and animal breeding

This is Juan Chavez again. I work for the Instituto Interamericano de Cooperacon para la Agricultura (IICA) in Peru and teach animal breeding at the National Agrarian University, La Molina.

Regarding message 40 (november 25) from Miguel Toro, I will ask the question: Is it not true that MAS gives more advantage when someone wants to identify with more precision a desirable allele? If not, then what are the advantages of MAS? I think with MAS the rest of the genome that does not code for the selected trait will remain less altered, compared to the less precise conventional artificial selection. Maybe the confusion arises when we do not take into account that mostly the responses of reproduction and fitness come from dominance and epistatic effects (combinatory effects) rather than to a specific allele. I understand this is the reason why, today, dairy cattle breeders are studying the responses of crossing dairy cattle breeds in order to improve the reproductive traits. Furthermore, regarding their precision, I believe MAS is closer to trangenesis than to classical artificial selection because with transgenesis it is possible to replace a specific gene without affecting the other parts of the genome. Maybe the problems of applying new technologies like MAS and transgenesis, which have to be better developed, is creating some "confounded effects" between what belongs to genetics and what belongs to the problems that arise from the application of the new tools of biotechnology.

Any type of artificial selection (focused on alleles coding for a specific response regarding a trait of interest) will affect fitness or the capacity of the population to be plastic to environmental changes. Genetic homeostasis and the limits of selection are close related with this underlying principle. I think fitness will always be reduced by artificial selection when traits other than reproductive traits are selected. Then, it is not necessary to study the detrimental effects of conventional or MAS artificial selection of a medium or a high heritable trait (i.e. fleece weight or live yearling weight) on reproductive performance, because we know, from a huge quantity of research, that this mostly always happen.

I agree that we Peruvians have to do research in order to identify the most valuable alleles, but that does not invalidate my point. I think any allele is valuable if it is different enough in its effects from others. And also, any allele with "detrimetal effects" in a given enviroment could have "selective advantages" in another environment. I am sure that our "Criollo" animals have valuable alleles because they live in very harsh conditions (high level of parasites, dry seasons with low nutritional inputs, freezing temperatures, low pressure of oxygen etc.). Regarding this, I am also sure that it could not be asserted that NATURAL SELECTION DOES NOT MAKE THE DIFFERENCE between almost 500 years of adaptation of Criollo livestock to the highlands of Peru (3000-5000 metres above sea level), since the sheep, goats, horses, cattle, poultry and pigs were brought by the Spaniards to our country. Finally, I think, any allele, any race, any species is valuable, more if we do not know what would be the future demands of we human beings.

Juan F. Chavez
IICA-Peru
Professor of Animal Breeding of National Agrarian University at La Molina
Lima
Peru
chavram (at) terra.com.pe

-----Original Message-----
From: Biotech-Mod2
Sent: 26 November 2003 13:06
To: 'biotech-room2@mailserv.fao.org'
Subject: 44: Marker assisted backcrossing

This is in response to message 42 (November 26) from Prof H. Dulieu.

Having experience with MAS, I strongly feel that marker assisted backcrossing for transferring one or more genes to a widely adapted variety, may not require larger number of backcrosses. Anything beyond two backcrosses may not stay economical and competitive. Our experience shows that selection based only on markers may not help in recovering the recurrent genotype. We suggest growing a larger BC1 population, rejecting about 50-60% of plants on a visual basis and analyzing the remaining 40-50% plants for markers and selecting 1-2 plants (with the desired marker and phenotype more closer to the recurrent parent) and backcrossing it to the recurrent parent. The same strategy can be followed in BC2. Again, growing a large BC2F2 population is important, wherein visual selection should be practiced prior to marker analysis. This will save a lot of time.

Dr Kuldeep Singh
Molecular Geneticist
Dept. Genetics and Biotechnlogy
Punjab Agricultural University
Ludhiana 141 004,
India
Tel. +91-161-243 30 81 (R)
kuldeep35 (at) yahoo.com

[Some definitions: Crossing 2 parents (P1 and P2) results in the F1 generation. Crossing F1 with P2 produces BC1 (also called BC1F1). Crossing BC1 with P2 results in the BC2 generation (also called BC2F1). Crossing BC2 with BC2 produces the BC2F2 generation...Moderator].