FAO-APHCA-AGAP
Workshop on Small-scale Dairying (SSD)

- Breeding for SSD -

Prof. Dr. Leo Dempfle

Technical University Munich, Germany

Dairy Production System

• Animal Production (dairy production) - complicated system

• Breeds/ Genetic improvement is one of several components

• What is best depends on the whole system!

You cannot optimize it in isolation!!!

• There are many factors and there are many interactions!!

• Since it is my topic

  I concentrate on the question of

  Breeds - Cross breeding - Genetic Improvement

How to approach 'improvement'??

• Take what you have as baseline (benchmark) and try to optimize (in an economic sense) the management

• Answer the question whether the conditions are good enough that other breeds might be feasible?

Can you import genetic improvement?

Breed comparison

Take into account all aspects which have impact on the economics

• milk yield (fat + protein yield)

• carcass value

• mortality

• fertility/reproduction

• health/disease

• etc. etc.

  The comparison has to be done
  at the same time under the same circumstances
  (good production circumstances)

  H_i = a₁g_1i + a₂g_2i +... + a_kg_ki

Difficulties of comparison

Having n unrelated cows of each breed

If we look at the confidence interval, we obtain

If the n cows are related (daughters of s sires, sk=n) then the accuracy can be much lower

Possibilities of exploiting breeds

One can use

1) an indigenous breed

2) an exotic breed (breed replacement)

3) (continuous) F₁ - Crosses

4) Criss-Cross breeding

5) Synthetic breed (varying proportion depending on...)

In order to decide you need

information, information, information!!!

116000 cows

58000 cows annually mated

2000 cows mated
to exotic bulls

56000 cows mated to N'Dama bulls

80% conception

800 males 800 females

640 heifers

1920 cows

2 880 000 kg milk

For 144 000 people

Forming a Synthetic breed

Forming a Synthetic breed

Using Criss-Cross(Rotational Cross-Breeding)

If you have decided for

• indigenous breed

• [exotic breed]

• Criss-Cross

• synthetic breed

you want to increase the efficiency further by

genetic improvement

of the genetic material (breed) involved.

However, genetic improvement

• is very slow and small

• and it is difficult to recover the expenses

but it is

• permanent

• cumulative

• multiplicative

The disadvantages:

Slow:

Improvement is often approximately of about one percent of the mean per year

Recovery of expenses:

There is usually no patent; once you sell an animal, the new owner can use it in his breeding - in his genetic improvement scheme.

But if you cannot recover the expense why doing (privately) genetic improvement work?

We have the situation that
a private good is becoming rapidly a common good

The advantages:
It is permanent and it is cumulative

Due to the present efforts (performance recording, intense selection) we achieve genetic progress which ins permanent.

The worth of the investment of a single cycle of selection is:

Repeating the cycle we add the new progress to the old one (it is cumulative)

It is multiplicative

Genetic improvement can be done in a relatively small subpopulation - the nucleus

In pig breeding
nucleus -several lines -altogether 1000 - 2000 sows
commercial unit - 8 000 000 fattening pigs

Nukleus

Work should be concentrated on the nucleus

Only there it is cumulative
(In the multiplier it is always diluted)

We have N animals tested each generation
We select K animals (cull N-K animals)

In the nucleus we must most accurately record

• performance

• pedigree (sire and dam)

The genetic progress per year is given

In each path we have i, r, t which vary

i: intensity of selection (K out of N); e.g.

1 out of 400: i =2.968
2 out of 400: i =2.813
10 out of 400: i =2.317
50 out of 400: i =1.642
100 out of 400: i =1.268
200 out of 400: i =0.796
400 out of 400: i =0.000

r: correlation between true and estimated breeding value

It depends on

• amount of information (own performance, dam's performance, offspring's performance)

• statistical procedure (BLUP)

• design of performance testing

t: generation interval (average age of parents at birth of offspring)

This is drastically different for a young sire scheme and for a progeny testing scheme

The expression DH/Dt has to be maximized for the given economic situation