This is just a reminder and a placeholder for a project that needs to get implemented but for which there is currently no time.
The marker effect estimation and the prediction of dgzw/gozw both use traditional ebvs (P or ITB) as input. The paths to these ebv files are stored in input files. Because the number of input paths in these files is quite large, a manual compilation of these lists takes a lot of time. As a consequence of that the compilation process should be done automatically.
Marker effect estimation and prediction have different inputs. These inputs are shown in the following table.
| Evaluation | EBV Type | Input |
|---|---|---|
| marker effect estimation | P | ebvFiles_P.csv |
| marker effect estimation | ITB | ebvFiles_ITB.csv |
| prediction | P | ebvFiles.csv |
| prediction | ITB | ebvFiles.csv |
From the above table one could suggest to just combine the two input
files from the marker effect estimation to the single input file of the
prediction. The problem with this is that in the set of breeds that are
considered is different in both processes. As an example, in the
prediction breeds such as rh, sf and
si are treated as separate breeds. Those are together all
as rh in the marker estimation.
Two possible ideas on how to implement the automated compilation of ebvFiles.csv can be thought of
Create a file ebvFiles.csv.template which contains all
the constant paths to ebv result files. The variable parts consisting of
the paths to ebv-files that change with every routine, are represented
with placeholders. These placeholders are then replaced by the actual
values.
Latest Changes: 2022-06-27 14:09:42 (pvr)