Successful performance recording starts with collecting quality performance information (raw data). This is the foundation from which quality EBVs are produced. Poor quality raw data is the source of poor quality EBVs leading to EBV queries, much breeder frustration and loss of confidence in Breedplan. The simple message is that if you present 'rubbish' to the computer for genetic evaluation, it will deliver you 'rubbish' EBVs in return.
How then can you improve the quality of the raw data that you collect?
The concept of evaluating the genetics of animals (producing EBVs) is based on head-to-head performance comparisons of individuals within a group, which are exposed to exactly the same conditions of, for example, feeding, animal health inputs, sex, status etc. By ensuring that these conditions are kept the same, the non-genetic source causing differences in production between animals in the group is standardized, thereby leaving the animals' genetic makeup as being the only remaining reason for the production differences.
If a group of animals is incorrectly managed, and the above does not occur, the quality of the raw data collected on these animals will be poor and the resulting EBVs will be of poor quality. This may lead to the wrong animals being selected for a breeding programme and disappointment in the way progeny of these animals perform.
The computer cannot sort animals into groups, unless it is told to, so in indicating different groups on your data entry forms, you are acting as its eyes. Consequently your decision, on which individuals go into which management groups, will have a direct impact upon the quality of the EBVs. While it is important to place animals in their correct groups, creating an unnecessary number of groups can also affect the quality of the EBVs. What starts out as a large group of calves can easily be broken down into a number of groups, the sizes of which are so small that the resultant EBVs are less reliable.
In addition to your grouping of animals on data entry forms, the computer automatically forms groups based on:
> The herd prefix
> The calving year
> The sex of the calf
> Whether the calf is a twin or a single
> Whether the calf is naturally conceived or an embryo transplant
> The age of the dam
> The blood percentage or grade of the calf
Some unnecessary grouping of animals can be avoided by simply separating them according to the above. For example, if for management reasons you find it necessary to split your cowherd up after calving, you can avoid forming a new management group by sorting cows according to any of the following:
> The sex of calf they are rearing
> The age of the cow
> The grade of the calf
> Those rearing embryo transplant calves as opposed to those not.
The following strategies will help you to improve the quality of your EBVs:
> Keep groups as large as possible Ð try to avoid single-animal groups, as these provide no genetic information on the animals concerned.
> Try to avoid single-sire groups, for the same reason as above, but applicable to the sire.
> Don't create unnecessary groups.
> Weigh animals before dividing into groups. For example, if you intend to castrate some of your male calves, delay castration until after the first weighing, then form another group (a steer management group).
> Weigh all calves within a group on the same day and on the same scales.
> Try to maintain strong pedigree links between groups.
> Collect information on as many animals as possible, including culls, steers and dead calves at birth.
> Use more than one sire in each management group, otherwise no useful genetic information will be generated on the sire. Remember that heifers up to 3.5 years are treated as a separate group to the older cows.
> Avoid changing all sires within the same year.
> Avoid running the same cows on the same area and breeding them to the same bulls each year.
> Keeping your calving compact will increase the number of calves in each calving ('slice') group.
> If cows within a herd receive different treatments during pregnancy, which are likely to affect the birthweight, gestation length or calving ease when the calf is born, the calves born to cows from these different treatments should be grouped separately on the calf entry sheets.
> If you believe cow management during pregnancy is likely to affect the weaning weight of their calves, then different groups need to be created on the 200-Day Weight forms. For example, the calf of a sick cow, who doesn't milk very well as a result of her illness.
> Animals within a group receiving different treatment, which is likely to influence their performance, should be grouped separately. Some examples of this are:
- Sickness gives some calves a permanent setback
- Some animals are fed for a show or sale
- Some animals are fed grain, others only grass
- Animals reared in different paddocks in which feed is of different nutritional value.
- A bull has been fighting and has clearly lost weight prior to being weighed.
- Yearling bulls used as sires/not used as sires.
- Calves weighed straight off the paddock, compared with those weighed three hours or more later.
- Different stages of pregnancy for heifers. Try to weigh before joining.
> At a given age, do all the required performance measurements on the same day. For example, at the yearling age you may want measurements on weight, scrotal size, fat depth and eye muscle area.
> After joining, run cows in as large a mob as is practical, rotating them through paddocks, rather than keeping them in two or more mobs in separate paddocks.
> Use cull females with EBVs as recipient dams in your embryo transplant programme, in preference to those without EBVs.
> Overlap AI and natural mating dates.
> There are a number of ways in which genetic progress can be made within a herd. Collecting good quality raw data is one of these.
> Raw data is often overlooked as being a source of good/poor quality EBVs.
> EBV queries, breeder frustration and loss of confidence in the system can often be traced back to the quality of the raw data submitted.
While the goal should be to collect 'perfect' raw data, this is rarely possible, nevertheless B+ grade data is better than C.