Sequence of programs in BLUPF90 package

Typical sequences

For BLUP run using small or medium data set

renumf90 blupf90

For variance component estimation using AIREML

renumf90 airemlf90

For variance component estimation using a Gibbs sampler

renumf90 gibbs2f90 postgibbsf90

For large-scale genetic evaluation

renumf90 blup90iod accf90

For Manhattan plots

renumf90 blupf90 postGSf90

For output of genomic- and pedigree-based matrices

renumf90 preGSf90

For single-step or GBLUP genomic evaluation

renumf90 preGSf90 blupf90

For large-scale single-step or GBLUP genomic evaluation

renumf90 preGSf90 blup90iod

Is renumf90 always needed?

All above analyzes can avoid using renumf90 if the input data is renumbered consecutively from 1, as is usually the case with simulation data.

Use of RENUMF90 usually guarantees consistency of data, pedigree, parameter and genomic files for later programs. This program: - renumber effects consecutively, - prunes the pedigree and does some quality control, - with genomic option, reorders genomic IDs, - writes parameter file (renf90), data, pedigree, and possibly genomic files.

After renumf90, the parameter file (renf90) may have to be edited for models not directly supported by renumf90.

Genomic analyses

With genomic analyses, no changes to the sequence are required as the genomic preprocessor program called preGSf90 is run automatically by application programs. PreGSf90 needs to be run explicitly for testing, when there is need to output some files, and when application programs need to read files created by preGSF90 rather than being recreated every time when the applications programs are run.