Back to index.html.

What if I do not want to use RENUMF90?

Some users do not want to use RENUMF90. Common reasons for this are:

• The data are simulated and already renumbered.
• I do my own renumbering and want to use it.
• I have a complex model and want better control over the assigned numbers.

In the following section, we provide instructions on how the files should be coded. We will not provide software, but you may take a look here.

Main requirements for recoding for BLUPF90 programs

• Levels of effects should be coded with consecutive numbers. The order may be “natural” or not. For instance, if the years in the data are (2010, 2011, 2013, 2014), they may be renumbered as (1, 2, 3, 4) or (2, 1, 3, 4).
• The animal is another effect, so it must receive codes from 1 to the number of animals in the pedigree file (this is usually greater than the number of animals in the data file).
• The animal effect does not have to be renumbered in consecutive order (with parents before offspring), although that order is valid and may be used. For instance, the following renumbering is valid:

  A 0 0
  B 0 0
  D A B
  E A D
  F B E
  Z A B

could be recoded as

1 0 0 A 0 0
2 0 0 B 0 0
3 1 2 D A B
4 1 2 Z A B
5 1 3 E A D
6 2 5 F B E

(the last three columns are not needed but help the visualization).

Unknown parent groups

However, if unknown parent groups (UPGs) are present in the pedigree, things get more complicated. Suppose that we have \(n\) real animals and \(m\) UPGs. Animals have to be renumbered with codes from 1 to \(n\), and UPGs have to be renumbered with codes from \(n+1\) to \(n+m\). In addition, because they do not have ancestors, the UPGs must not have their own lines in the pedigree file for BLUPF90. This way, BLUPF90 “knows” that a given number is a UPG and not an individual. For instance, in the previous example, assume that the UPGs are “unknown2000” and “unknown2004”:

  A unknown2000 unknown2000
  B unknown2000 unknown2004
  D A B
  E A D
  F B E
  Z A B

It should be recoded as follows.

1 7 7 A unknown2000 unknown2000
2 7 8 B unknown2000 unknown2004
3 1 2 D A B
4 1 3 E A D
5 2 4 F B E
6 1 2 Z A B

Genotypes

Genotypes should be coded as 0/1/2 for {AA, Aa, aa} and 5 for missing, as described elsewhere in this tutorial, in a fixed-width text file. The first column of this file contains possibly alphanumeric identifiers for animals, for instance:

   A 120120202102111
   Z 121111202111010

The important requirement is the creation of the cross-reference file (usually called _XrefID). This file contains the original ID and the new ID in the renumbered pedigree and data files. In the previous example, it would be:

 1  A
 6  Z

Overall mean

BLUPF90 programs do not include an overall mean by default. If you need one, add it yourself as a column of 1s in the data file.

Back to index.html.